Skip to content

The Fruitless Quest for Digital Media ROI

2009 October 8
by Greg Satell

There is a lot of misguided talk about ROI lately and it will probably do more harm than good.  Most of the efforts are well meaning, yet they enlighten more about the immaturity of the Digital Media industry than they do anything else.

The errors range from a misreading of the term to a lack of understanding about how advertising works to naïveté that manifests itself in suppliers believing everything that clients tell them.

I will try to dispel some of the confusion.

What is ROI?

ROI, for those who don’t already know, is a financial term and stands for “Return on Investment.”  Although recent events have shown Financial Industry metrics to be found wanting as well, Wall Street types make a lot of money and so are assumed to be smarter than the rest of us.  Therefore, it’s not surprising that marketing people want to emulate them.

However, if one looks at the way financial people use ROI, the error should become immediately apparent.  ROI was never intended to be used for every, or even for most investments. Its main function is to evaluate businesses as whole or major investments such as an entire factory or business components that have a major effect on the bottom line.

Most investments, such as office furniture and equipment, employee training, etc. are not broken out separately.  Even many high ticket items, ranging to specific machinery to multimillion dollar art in CEO offices, are not calculated.  They are simply aggregated into larger investments.

ROI vs. Accountability

Accountability, however, is important.  Suppliers have a responsibility to deliver what they promise and buyers have a reasonable expectation to get what they pay for. Like any contractual relationship, the metrics are decided by mutual consent.

Caveat Emptor cautions the consumer to beware, but the same should go for suppliers as well.  It is all too easy for buyers to demand and all too tempting for suppliers to promise.  Often, naïve media people are taken in by buyers who have a lot more experience and information on their side.

Accountability is, for the most part, built into media transactions: CPP, CPM, CPA, etc. Suppliers such as Publishers and Digital Media Agencies agree to these metrics when they sell their wares.  The trouble starts when buyers get suppliers to agree to metrics that they can’t possibly deliver on effectively, such as the sales effect of a specific media or marketing action in an overall marketing effort.

Does the Difference Matter?

I would vigorously argue that there is more than semantics at stake here.  Many Digital suppliers completely misread their role in the process.  A good example is this presentation.  Although entertaining and well done, the guy is not only headed for trouble, but incites others to run off the cliff with him!

He urges Social Marketing people to find a correlation between their activities and sales.  It seems reasonable on the surface, yet only if one assumes that Social Media constitutes the bulk of marketing activity.  Moreover, the subtext is troubling.  He actually encourages confirmation bias.

The flaw in logical reasoning would be the same in finance as in marketing.  Every investment simply can not be accounted for.  Just as nobody calculates ROI for chairs and copy machines, nobody should realistically expect a direct correlation between every marketing action and sales.

Digital Media accounts for about 10% of global ad spend.  Even in high water markets, like the UK, shares are only in the mid 20’s,.  In all but the most specific cases, investment in Digital as a whole, much less any specific action, is dwarfed by other activity that is integrated into an overall marketing program.

An ROI calculation is just not feasible.  Which is why, for most marketers, Digital ROI is irrelevant.

Moreover, advertising isn’t the only thing that affects sales.  There is the comparative quality of the product, distribution, service, competitive activity, etc. Advertising has to do its job, but so does everything else.

It behooves the Digital industry to grow up if profits are ever to come in line with valuations.  It is easy to blame clients who sometimes demand things from suppliers that they neither want nor need, but that’s misplaced.  If one is stupid enough to agree to what one shouldn’t, the blame lies with oneself.

Awareness vs. Direct Response

I believe that the problem stems from the foolish overemphasis on direct response in the early days of the internet.  While the direct response model can be very effective for some clients, for most it is not a full solution.

A good example is the auto category.  Most car ads are not supposed to result in a sale.  People don’t buy cars very often and so they put a lot of time and effort into researching the purchase.

Nobody can investigate every possibility so it is understandable that consumers only consider a few brands.  Marketers call this phenomenon the “consideration phase” and for many products, it last a lot longer than the active buying phase.

Much if not most of auto marketing budgets goes toward image advertising that encourages consumers to consider their brands.  Although some expenditure goes toward direct response as well, consumers are unlikely to respond to offers for brands that they are not considering and as noted above, the consideration phase lasts longer.

The average consumer is only capable of keeping 3-5 brands per category in their working memory, even those that are heavily marketed. (Try naming car brands off the top of your head and see where it starts becoming difficult).  It’s a zero sum game.  If you want to get your brand on a consumer’s consideration list, you have to knock another brand off.

That’s why brand awareness is so important.  People generally don’t consider brands they are not aware of unless there is a deep discount or other special offer, which can be more expensive than advertising.

Can Marketing ROI be Measured?

Theoretically, it is possible to measure marketing ROI, but only under certain conditions and even then the results aren’t entirely conclusive.  Even companies who have had success in this area are somewhat of a cautionary tale.

Research companies like Millward Brown do offer brand tracking services to their clients.  However, most companies simply can’t afford the expense.  ZenithOptimedia has had great success since it repositioned itself as the “ROI Agency.”  There are other success stories as well.

However, successful ROI efforts come from companies who have very close and longstanding working relationships with the clients they serve and enjoy access to confidential internal data.

Moreover, these companies constantly struggle to manage expectations and deliver on them.  In effect, they are successful because they have unique access and keep the definition of ROI fairly narrow.

A more sophisticated and expensive option is econometric modeling.  There are a variety of companies that offer the services of PhD level mathematicians who build bespoke multivariate models for clients.  I myself have had some limited success with two-variable models in excel.

However, even the most sophisticated econometrics doesn’t explain all of the variability in real world data.  Definitive causality is a chimera, a pipe dream and a sham.

What is the ROI Answer?

Common sense goes a long way.  Don’t promise anything you can’t deliver.  Don’t agree to what you don’t understand.  Try to do a good job.

Just because some clients demand the world, doesn’t mean that they actually expect it.  If you go ahead and make a promise every time a client raises an issue, you’ve got much bigger problems than ROI.

– Greg

35 Responses leave one →
  1. October 8, 2009

    This is all very high level, Greg. It’s sophisticated business, but most people are not that sophisticated. What about small-business owners? They can’t hire advanced agencies to model ROI, so how do they know if advertising is working if they’re in an industry that doesn’t respond well to direct marketing?

  2. October 8, 2009


    The advantage that small businesses have is that they have a lot less activity and there is a lot less noise in the data. As I said in the post, I don’t think ROI is realistic unless the marketing program is very small.

    Insight, however is achievable through tracking. One very simple thing to do is to track marketing inputs (i.e. activity) and marketing outputs (i.e. sales, leads, etc) in an excel sheet. After a few months, you should have enough data to do some simple two variable models and you can get some idea what drives outputs.

    You have to be careful though. Once, I had a consulting client who thought their advertising wasn’t working and wanted to cut their budget. What was actually happening was that their competitors had increased their spending and now my client’s ads were getting drowned out. (The econometrics were incredibly clear on this point. I don’t think I’ve ever seen a cleaner model). What they actually needed to do was redouble their efforts.

    Data is an input, not an output.

    – Greg

  3. October 9, 2009

    Thanks for the insightful article Greg! As someone who is a Service Provider to corporate houses in the Digital Marketing space (Print & Email), your arguments will help me in addressing the common objections that one faces.

  4. October 9, 2009


    Thanks. I’m glad you found it helpful.

    – Greg

  5. October 9, 2009

    Greg, there are some fair points to consider, but the overall tone of this article referring to ROI as a “fruitless quest” is something I totally disagree with. Some particularly troubling statements:

    “Just as nobody calculates ROI for chairs and copy machines, nobody should realistically expect a direct correlation between every marketing action and sales.”

    “An ROI calculation is just not feasible. Which is why, for most marketers, Digital ROI is irrelevant.”

    Social media costs are more akin to marketing and advertising costs than social media consultants want to admit, and if not successful, will be axed (just like an ad campaign that’s doing nothing.) Your example of someone’s ads being drowned out is a good insight, but doesn’t change an ROI argument for advertising. You still have to show that doubling the ad cost is going to be effective and ultimately, there will be some return for all the money spent.

    A chair for someone’s behind to sit in, is a lot less than the $50,000 to $100,000 annual cost for an employee or consultant who is spending all day at Twitter & Facebook talking to prospects and customers, among other socially-related activities. Social media results can be hard to measure like ad campaigns, because as you mention there are factors other than advertising that result in purchase. Plus, you may not be positive the traction and momentum is coming from social efforts or somewhere else, but that is precisely WHY the man you refer to from South Carolina recommends doing timelines of activities (social, marketing, advertising, charitable, etc.) and then looking BACKWARDS after spending some time collecting this data, you ought to reasonably see results in various places, revenue being one of them. It may not be as concise as you feel it should be, but it is an approach that I believe is very helpful and helps put some context around the “vagueness” social media folks seem to be so fine with. If you actually tried this method (you’d have to give it a time period of six months or more to really gauge it) I know there would be some revealing, helpful data to analyze, and areas to consider changing or improving. That’s not a fruitless endeavor nor a waste of time.

    Come on… you expect jobs to be created and consultant “promises” enough, without some attempt to show financial value and monetary gain to executives who ask for it? Like it or not, at the end of the day, all these companies have to make money to stay in business. Otherwise, social channels are totally irrelevant to them.

  6. October 9, 2009


    Thank you for your thoughtful comment.

    I think you might have misunderstood. Let me try and clarify:

    Chairs, copy machines etc: Social Media costs ARE marketing costs, they just make up a very small part of marketing costs. Moreover, their effects are intertwined with other marketing and service efforts. Response for social media, and almost any specific media action (unless it is very, very big) is going to be highly dependent on other actions.

    Digital ROI is irrelevant: That statement linked to an article about a study that found that while 76% of advertisers did evaluate traditional advertising, only 47% conducted reviews for their Digital marketing because either spending was too small, they were too busy or they didn’t get around to it.

    So if advertisers themselves don’t think it’s important to evaluate their entire Digital budget themselves, why do so many Digital Media suppliers feel that ROI is so crucial to deliver ROI metrics for their small part?

    I do agree that response should be evaluated, but response and ROI are two different things. Unless you plan to work with very small marketing budgets where social media dominates, ROI is really a waste of time. As you noticed in the South Carolina presentation, his “ROI metrics” didn’t include any other factor except for Social Media.

    I also agree that monitoring and analyzing response can yield important insights. Yet, again, insights are also not ROI. It’s doing a very good job for your client, but it’s not ROI.

    If you were an integrated marketer who might work with dozens of suppliers, would you accept dozens of ROI calculations or would you evaluate your marketing budget as a whole? If evaluating ROI was as simple as measuring response, why would major marketers spend so much money on tracking research and econometric modeling?

    From the passion you show in your comment, I’m sure you do a good job and deliver tremendous value to your clients. However, delivering value and being responsible for ROI are two different things.

    Again, thanks for stopping by and commenting.

    – Greg

  7. October 10, 2009

    Hi Greg. I enjoyed reading your post today and added you to my reading list. Let me just say a few things first.

    One: You’re entitled to your opinion, and I completely respect that. 😉

    Two: Though I agree with you on some points, I disagree on several. Here they are:

    1. You state that “ROI was never intended to be used for every, or even for most investments. Its main function is to evaluate businesses as whole or major investments such as an entire factory or business components that have a major effect on the bottom line.” I disagree.

    Whether or not you want to worry about ROI is absolutely up to you as a manager or business owner. No argument there. You don’t necessarily want to try and calculate the ROI of a cocktail party or the ROI of a customer service department, or even the ROI of a billboard. I get that. But that is a matter of operational choice and a far cry from the notion that “ROI was never intended to be used for every investment.”

    The reality of ROI is that it can be used to determine the value of any and every investment an individual or company makes. When it comes to ROI, the intent of the equation isn’t an issue: Even if that statement were accurate, even if whomever devised the ROI equation didn’t intend for it to be used for certain activities and actually stated that (can you point us to something along those lines?), it has no bearing on whether or not it should be used where it can be used.

    ROI can be used to determine whether or not ANY investment resulted in a positive return, and to what extent. The medium, the context, the reason… all of these are up to the “investor” to decide.

    2. “The trouble starts when buyers get suppliers to agree to metrics that they can’t possibly deliver on effectively, such as the sales effect of a specific media or marketing action in an overall marketing effort.”

    In terms of predictive modeling and estimated success, I agree 100%.

    However, you can measure the effectiveness of certain activities pretty easily in a variety of ways. Some will lead to transactions. Others will not. Different places for different types of impact measurement, from web traffic to brick & mortar traffic to online mentions, sentiment and awareness polls. You can also seed promo codes specific to each channel (even to each billboard) to track the direct impact of every media outlet on actual transactions.

    3. I am not a poor misguided fellow from South Carolina. I’m actually a poor misguided fellow from France who happens to live in South Carolina. I think we can dispense with the pejorative characterization though. But hey, to each his own.

    4. You write: “It seems reasonable on the surface, yet only if one assumes that Social Media constitutes the bulk of marketing activity.” That is absolutely incorrect.

    I don’t make that assumption at all. As a matter of fact, I make no assumptions. Ever.

    An assumption usually looks like this:

    “ROI was never intended to be used for every, or even for most investments. Its main function is to evaluate businesses as whole or major investments such as an entire factory or business components that have a major effect on the bottom line.”

    Or even this: “It seems reasonable on the surface, yet only if one assumes that Social Media constitutes the bulk of marketing activity.”

    You’re letting your own bias and resulting assumptions about what I am talking about send you down the wrong path on that point.

    The purpose of the presentation is to simplify the concept and general methodology behind ROI measurement in SM and make it easy to grasp. I purposely isolated SM activities to keep things from getting confusing. I also had less than an hour to present this, so I have to keep things on target. I could elaborate on measurement across all channels media and departments, but can’t fit that in a one hour presentation.

    You should note that several of the graphs I use as examples incorporate a number of activities that have absolutely nothing to do with SM. I did this to illustrate that this is not limited to SM. If you missed those, take a closer look.

    5. I absolutely do not “actually” encourage confirmation bias at all. That is completely incorrect. Again, you need to go through the presentation again. What I recommend is looking for correlations first, and then trying to disprove cause and effect.

    This is important, so I will repeat it again: Once you have discovered a possible correlation between an action (or series of actions) and a reaction, you want to disprove cause and effect, not assume cause and effect.

    Only once you have run out of ways to disprove a connection can you start the process of trying to truly establishing a connection between action and reaction. It’s basically the scientific process applied to measurement and business performance analysis.

    I am not sure how you missed that.

    6. You also write “An ROI calculation is just not feasible. Which is why, for most marketers, Digital ROI is irrelevant.” Wrong.

    Just because you don’t understand how to do this or because you’ve been burned by people who claimed to know how to do it but didn’t doesn’t mean it can’t be done. Or that it’s irrelevant. The two statements contained in that passage are assumptions, and they are 100% incorrect.

    a) ROI calculation is feasible. You just have to know how to do it.
    b) What is relevant (or convenient, rather) to marketers isn’t the issue. What IS the issue is what is relevant to the managers at the companies investing $x into activities. And what is relevant to them is knowing where their budget dollars are scoring and where they are flopping.

    Frankly, I couldn’t care less what measurements are relevant to a service provider or contractor. As the client, I’ll tell them what’s relevant to me and my business.

    If businesses relied solely on what marketers think is relevant, they would measure success in web traffic, click-throughs and impressions. Big deal. So 1000 new people visited your website this month. How many bought something? How many plan to buy something? How has this impacted my bottom-line or my market share?

    Know what I mean? If you can’t tie this stuff to business performance, you might as well be measuring how many times people blinked while checking out your billboard on their way to work.

    7. You also seem to imply that Social Media is mostly a digital marketing medium. It is not. Social Media is increasingly being used to improve customer service, customer relations, collaboration and conduct market research.

    I won’t assume that you can’t see past this point (again, unlike you, I don’t make these kinds of assumptions), but in case you do, please consider that the “direct sales” conversation only applies marginally here. There’s a bigger picture to how Social Media integrates with brand management, brand perception, user experience, etc. Don’t look at it as mostly another set of push channels. It isn’t a new form of digital/conversational advertising hybrid.

    8. “Most car ads are not supposed to result in a sale.” Seriously? You’re kidding, right?

    9. “Even the most sophisticated econometrics doesn’t explain all of the variability in real world data. Definitive causality is a chimera, a pipe dream and a sham.” This may come as a surprise to you, but I agree with you 100%.

    I have worked with two such companies. Both promised to deliver predictive modeling. Neither delivered. (I suspected as much, but my bosses at the time really believed in the black magic these PhD guys from MIT were selling.)

    I played along, keeping an open mind, but my suspicions were on target. You just cannot accurately predict consumer behavior based on pre-existing and constantly shifting conditions. More importantly, the cost of running such predictive analysis programs far outweighed the “estimated” financial gains.

    In other words, running a simple ROI analysis of this type of program identified it as a fumble to senior management. Even though you advocate that ROI analysis probably shouldn’t be used for this type of application, it was the only thing that convinced senior management that they were wasting money on these charlatans.

    But back to the point: While trying to predict causality and influence reaction though specific action based on that predictive modeling is limited at best, you can measure that relationship in real time by looking at actual data (not estimated data).

    Predictive modeling may be a sham, but marketing mix analysis is not. Don’t confuse the two. One uses BS mathematical models while the other uses real data, business savvy and objective analysis.

    So Greg, I agree with a lot of the things you’ve stated here, and I think that you and I are on the same side of the fence. We don’t like the usual measurement and marketing BS we’ve been witnessing all our careers and it’s thick in the world of digital marketing and SM. BUT, you really need to do 2 things:

    1. Go through my presentation again and look for the nuggets you missed. I think you misunderstand the measurement philosophy I discuss in it. Better yet, go spend some time on my blog and you’ll probably start to see my position a little better.

    2. If there’s ever a chance for us to meet and discuss how this works, I can show you. I’ll be happy to answer all of your questions (if I can) and show you that I’m not falling prey to confirmation bias.

    If you can make it to London’s SMIB conference on the 23rd of this month, I’ll be there. We can chat then. (I’m not sure when I’ll be back in Europe anytime soon.)

    What’s really important here is that you realize that ROI calculation when it comes to Social Media, marketing, business development, etc. is both relevant and real. I know both points seem dubious to you right now, but you have to give me the benefit of the doubt on this. I’ve been working with this stuff for years, long before Social Media was a term, and I’ve pushed past mountains of nonsense, false assumptions, bad math, and poor science to find what works and what doesn’t. Most of it is simple common sense, attention to detail and hard work.


    Olivier “the poor misguided fellow from South Carolina” Blanchard

  8. October 10, 2009


    Thank you for coming and taking the time to write such a thoughtful comment. I will refer to your points in the same order:

    1. ROI should not be used for every investment: I don’t think it’s actually important what the investment is, but what contribution it makes to the overall business. In some cases, Digital Media, Social Media or any other marketing action can make up the majority of investment, and if it does, ROI can be calculated. However, that is a minuscule minority of marketers, so I stand by the point.

    If a marketer had a heavy TV campaign that raised spontaneous brand awareness by 10 points, wouldn’t that affect the ROI calculation? It doesn’t in your presentation.

    Conversely, let’s say that because of the economic crisis a marketer cut all of their ATL investment and switched to low budget social media . Most likely, response would decrease. Should the Social Media person be blamed?

    In either case, to make sense of the data some modeling would be needed. I think making a distinction between response and ROI is important.

    2. If you are talking about measuring response, I agree. However, for an integrated marketing program, all of the components need to be taken into account to measure ROI. ( i.e. We spent “A,” competitors spent “B” and our market share changed “C.”).

    Predictive modeling is the only way that I know of to parse the components. However, predictive modeling never explains 100% of the variation in the data (r-squared). So insights can be gained, but definitive causality is not realistic.

    As a guy who runs an econometric division for one of the global ad networks told me once, “data is an input, not an output.” I think that’s good advice.

    3. Pejorative descriptor: You’re right. Sorry. I will change the descriptor. Viva la France!

    4. I looked through the presentation again and I still don’t see it. There were some things on a few slides (I saw a conference) that weren’t strictly social media, but it still seemed dominated by Social Media. I realize that there are limits to what you can present and you might have had something more in mind while you were writing it. However, the purpose of a presentation is to communicate and that’s what was communicated to me.

    Nevertheless, a lot was left out. You show no competitive activity, for instance, which can have an enormous effect on campaign performance. Furthermore, once you start calculating ATL activity, in-store promotions, competitive monitoring, etc. you are no longer talking to Social Media Managers but Communication Planners.

    5. Confirmation bias: Confirmation bias occurs when the one doing the testing has an interest in the results of the test. If that is not what you were encouraging, please explain. As I understood the main theme of your presentation was that Social Media people should evaluate Social Media campaigns.

    6. For most marketers, Digital ROI is irrelevant: That statement linked to an article about a survey of marketers which showed that a majority (albeit slight) don’t even do performance reviews of their Digital Agencies. The most common reason given was that they don’t spend enough.

    Maybe it is relevant to you as a client and I’m sure it is too many others as well. However, for a majority, it isn’t.

    As to your comment about marketers, I would say that it depends on the marketer. Like any other profession, some are better than others. However, a good marketer cares about sales.

    7. Social Media is not just marketing: I agree, but don’t see how that makes measuring ROI more feasible. In fact, it seems to me that it makes measuring the benefits of ROI less feasible. It is very difficult to measure the profit impact of a happy customer.

    8. Most car ads are not intended to result in a sale: No. I’m not kidding. Try to model a car ad budget against sales on a time series. It’s impossible. They spend huge amounts in a launch year and make their profits in the remainder of the model life. I challenge you to find me one national car budget that doesn’t work that way.

    Of course, dealership budgets do focus on direct response and I mentioned that direct response ads are important as well, but they don’t tell the whole story.

    9. I hate to disagree when you actually agree with me, but I have to. The econometrics people that I have met have been highly thoughtful and I have learned a lot from them. One of the benefits is that a model shows not only the direction, but the strength of the relationship. While there is always variability that can not be explained, powerful insights can be obtained through good modeling. Actually another problem I had with your analysis is that it was a bit too graphical, which can be biased by scale.

    While I do agree the rigorous approach you suggest to Social Media is valuable and I’m sure that you’re good at your job, I really have a problem with the whole concept of Digital ROI. As I stated in the article, Digital accounts for only 10% of advertising budgets globally. The other 90% is bound to have a greater effect.

    While I agree that Social Media is valuable and worthwhile, I think you could be doing a greater service by putting Social Media in it’s proper perspective. At the moment, Social Media Marketing is growing while overall marketing budgets are being cut. The bottom has fallen out of sales for most industries. Should Social Media be blamed? Of course not!

    Yet, according to your presentation Social Media should be. Social Media is up, overall sales are down.

    My overall point, and maybe I wasn’t clear enough, is that I strongly believe that a distinction needs to be made between measuring response, accountability and ROI.

    Again, thank you for taking the time to write such a thoughtful response.

    – Greg

  9. October 10, 2009

    I can see how this discussion could go on for a while. 😀

    1. I hear ya, but you’re making an assumption I’ve noticed before: That somehow, spend has a proportional return in terms of effect. (Low spend, low impact. High spend, high impact.) No. You could spend 80% of your marketing budget on advertising and only get a 4% return (about the average by some studies) while spending only 5% of your budget on the social web an see 20+% of your sales driven from that social engagement.

    A client of mine at a previous company realized, after this kind of analysis, that while his print advertising increased awareness of his stores and sales, it drove neither traffic nor sales. However, email marketing and Facebook drove over 30% of his brick & mortar traffic on the weekend, and over 50% of his traffic on days when he had a special promo or sale.

    Most of his marketing spend was in print ads. It generated almost zero traffic or sales dollars. His email and FaceBook activities, which cost almost nothing to produce generated a huge chunk of his traffic and sales. A lot more went into this analysis, but to make a long story short, he stopped wasting his money on print ads.


    A) Where you spend the biggest chunk of your budget has no bearing on where you should look for ROI.
    B) The biggest item on your budget may not have the biggest impact on your business (and probably doesn’t). I would seriously caution against that kind of assumption.

    To answer your question about changing economic conditions and the TV ad awareness thing: Adjustments do need to be made for changing conditions, of course. If consumer spending is down across the industry, you account for that. Market share comes into the analysis, partly to make sure that numbers aren’t misinterpreted. It isn’t in the presentation but it came up as a question after the session. Remember that the deck is a very basic overview. It isn’t the training course. I can’t cover every detail in an hour. 😉

    Incidentally, a TV campaign would go into the activities timeline and resulting changes in awareness, mention, sentiment and activity would, yes, be looked at as possible precursors to transactions.

    2. Data is data. I’m not sure where input or output even come into that conversation.

    That said, I don’t have a high opinion of predictive modeling. I’ve worked with 2 well known firms that claim to have PM figured out, and they both sucked. (They couldn’t predict their way out of a paper bag if their 6-figure fees drew them a map on the backs of the checks we wrote them.)

    It isn’t to say that someone, somewhere has the predictive modeling thing figured out, but the ones who work with some of the world’s biggest brands sure as hell don’t.

    Incidentally, MPI (market performance indicators) and net KPI are completely different things. Market share and net sales, for example, don’t live in the same house. You can look at the two side by side to see if your net growth is a result of your activities rather than a market trend you benefited from, but that’s about it. Bringing a competitor’s spend into the ROI equation doesn’t make a whole lot of sense to me. But… maybe I misunderstand what you’re talking about. If so, my apologies.

    3. We’re good.

    4. A lot was left out of the presentation, yes. Reasons: Time constraints, purpose of the presentation, and the need to keep things as simple as possible. Also bear in mind that the slide deck doesn’t include my commentary. Many of your questions are probably answered by my actual presentation, not just the slides I flash in the background.

    If I were to explain how the whole thing works (which works best as hands-on training), we would be looking at thousands of slides and probably a week or two of sessions.

    Not that it’s rocket science, but there are lots of moving parts and it takes time to cover them all and help map it all out for someone who is new to this.

    5. I’ve already answered that: a) Look for correlations/patterns. b) Disprove causation. c) Whatever is left, see if you can actually prove causation. It’s a process of elimination. Though you want to look for patterns initially, what comes afterward is an objective process of elimination. There is absolutely no room for confirmation bias in my world. None. Zipola.

    6. Marketers who don’t bother to audit or assess the performance of their digital departments or agencies are negligent in that respect. Just because the majority is asleep at the wheel (or doesn’t know how to measure success in that space) doesn’t mean we should all follow suit.

    7. It’s simple: ROI lives in $$$. That’s it. What’s the value impact of a happy customer? In terms of ROI specifically, it can only be one of these things:

    a) Increased frequency of transactions. (Starts buying more often.)
    b) Increased depth of transactions. (Start buying deeper into the product line.)
    c) Increased yield of transactions. (Starts spending more per transaction.)

    That’s it.

    Everything else (sentiment, mentions, recommendations, positive opinion, increased awareness, etc.) is non-financial and therefore not part of the ROI discussion. These may be precursors to ROI, they’re great KPIs, they matter a lot, but that’s about it. While most marketing and media measurement folks love to quantify that kind of thing and pass it off as ROI, I look at it all as sets of intermediate data.

    10,000 extra visitors to your website, higher awareness, increases in mention and positive sentiment, etc. don’t amount to jack diddles until they turn into an increase in sales. Period.

    ROI is either a cost reduction or an increase in revenue. A happy customer who recommends you to ten friends and brings them to your store doesn’t become part of the ROI “equation” until some of those friends slap down some cash and buy something.

    That’s why the ROI discussion is not an SM discussion. It has a specific SM component, but it is only relevant to SM when SM somehow impacts sales (indirectly or not) or a cost reduction.

    Example of a cost reduction brought about by social media adoption: A customer service inbound call center decides to use Twitter to address some online complaints and questions. Ticket resolutions go from 20 minutes to 5 minutes. The company starts seeing a steady shift to Twitter by customers looking for quick customer service. A year later, the company can manage the same number of tickets it handled a year ago with 50% of the headcount.

    8. Just because it doesn’t work that way doesn’t mean that’s not what it is intended to do. These ads show the car, pimp the car, make you crave the car, and give you the price of the car and the sales incentives (financing, special promos, etc.) Of course the purpose of the ad is to sell the car. Dealers and automakers don’t spend millions on advertising just to raise awareness about their brand and models. They want to move steel, and preferably before the end of the month.

    9. I’ve already expressed how I feel about predictive modeling… or rather, the faith that some execs put in predictive modeling. Insights are nice. But insights aren’t predictive. So maybe we’re looking at a bit of a misnomer there.

    Companies hire PM firms to tell them what to do to achieve results XYZ in the future. PM firms tell them what they’ve done to achieve XYZ in the past, and mathematically try to project this into the future. Well, okay… assuming all things remain more and less the same, that would work fairly well. Unfortunately, conditions change, and they change fast. From the economy’s health to advances in technology, shifting consumer tastes and aggressive pricing tactics by competitors… Predictive modeling has the underpinnings of a science but the application is mostly alchemy. To make matters worse, it is prohibitively expensive. So… sure, I’ll have beers with those guys, but that’s about as far as it will go. I’d rather invest in a talented, enthusiastic staff than waste my money on trying to combine good math and flawed assumptions to tell me what tactics to favor in a hypothetical future.

    And there’s that notion of yours about the ratio of spend to impact/effect:

    You wrote “As I stated in the article, Digital accounts for only 10% of advertising budgets globally. The other 90% is bound to have a greater effect.”

    “Is bound to?” Nope. Spend and impact are not necessarily proportional at all. Those kinds of assumptions and generalizations will get you nowhere fast. Don’t fall for old advertising thinking. Sometimes, it indeed works that way, but it is far from being a rule, especially moving into the next decade.

    If anything, you can probably get a lot more bang for your buck by shifting your mix to digital (especially mobile) these days. Social lives on mobile platforms more and more.

    Not to mention that social isn’t digital. We’re talking about people connecting with people, here. The technology is the medium, not where the investment, tactics and impact actually live. 😉

  10. October 10, 2009


    Thanks again for responding. Here’s where I think we agree: Sales effects can be tracked and relationships between actions and relationships between actions and sales can be derived. However, that is not necessarily ROI.

    I have a few problems in the logic of what you said. Some examples:

    1. Aren’t you doing predictive modeling yourself? What would be the point of making so much effort to analyze the past if you assume no predictive value for the future?

    In your presentation you say “all you have to do is show what the revenue gains and cost savings are and plug them into the formula.” For what purpose? To show that you did a good job or to improve efficiency in the future?

    The only difference I can see between what you are doing and what econometricians do is that econometrics produces not only a model, but model fit (r-squared). They can say that “this model explains 60% of the variation in the data. There is something important in the factors measured, but still a bunch that can’t be explained.”

    To me, that’s a good thing. In my own ROI efforts being able to know when my model has stopped working has been invaluable. I really don’t understand why you have a problem with econometrics. Please explain.

    2. Aren’t you really talking about communications planning? Once you start putting TV on the timeline, you are no longer talking about Digital or Social Media ROI, but communications planning. You would have to know something about TV buying and planning to evaluate it effectively. If there was a lot of Print, you would have to know something about Print.

    3. You don’t define ROI: As you yourself pointed out, my definition is clear. What marketing outputs you get for given inputs. The goal of the insights gained is to make future investments more efficient.

    This point is very important as any Communications Planners with a ROI contract will tell you. Defining ROI is a major negotiation point.

    4. Why do you think competitive activity wouldn’t affect sales? My own experience tells me different and the amount of time and effort that Communications Planners spend on competitive analysis makes a strong case that this point is very well established.

    The same point would go for sales vs. market share. You talk about benchmarking, but market share is simply company sales benchmarked to category sales. If you don’t take market share into account then you are looking at sales in a vacuum. Marketers would be heroes in up markets, losers in down markets. That doesn’t seem like a good idea to me.

    5. You say, “Spend and impact are not necessarily proportional at all.” Isn’t that the point of ROI, to evaluate the efficiency of investment? If not, what is the point?

    My biggest problem is the difference between what you say and what you do. I really don’t have a problem with what you do. I’ve seen enough of your approach to believe that what you do is completely worthwhile.

    However, you seem to encourage everybody to follow the same approach and I think that is a very, very bad idea.

    1. Suppliers should not, as a rule, engage in ROI (although there are exceptions). I believe that the reason that my article has attracted such interest is that so many Digital suppliers believe that it is their responsibility to deliver ROI (or their clients are trying to make them believe that it is). That would assume that integrated marketers should give confidential internal data to dozens of suppliers and should accept dozens of ROI calculations back (all of which are bound to conflict).

    2. You continually say that your approach is successful, but predictive modeling is not, although you admit that many if not most of the world’s premier marketers spend enormous amounts of money on tracking and econometric modeling. I’m sure you mean well, as I’m sure that your passion makes you good at what you do. However, I think the case that you are right and everybody else is wrong is a hard one to make.

    I would submit that both approaches can be successful, but scale is important. As brands become bigger success factors multiply. Marketing Success often breeds failure because success attracts a more vigorous and complex competitive environment.

    3. You deny Confirmation Bias: The concept that someone interested in the outcome of an analysis would have no (zippo?) tendency to see the results somewhat differently than someone who is completely disinterested is a bit hard to take. Confirmation Bias can be minimized, but can never be eliminated.

    4. You don’t qualify your remarks: The fact is that the world is a messy place. There is always a lot that can not be explained. Sometimes you can find very clear relationships, but often you can’t.

    A good example is billboards, which defy precise ROI analysis. It’s clear that they work (and I have seen very impressive results for billboards on tracking studies), but nobody really knows how. Should marketers forgo using billboards just because a precise relationship can not be derived?

    Marketers engage in many activities where precise ROI can not be derived. In fact, they spend millions of dollars on such activities, just as corporations spend millions of dollars on employee training. No one doubts that these activities have benefits even if they can’t be measured precisely.

    Finally, everybody can not be responsible for ROI. An integrated marketing program includes ATL media, digital media, direct marketing, POS, Creative, Events, etc. Everybody needs to be responsible for their role in the process, but having everybody running around trying to calculate precisely their contribution would just create an enormous mess.

    That doesn’t mean that everybody shouldn’t make the effort to track response(again, I think the distinction between ROI and response is important). However, the biggest value in tracking comes from uncovering what you don’t understand. If a digital or direct campaign results in an unusual response, that’s a signal that something else is going on. Maybe there was a TV campaign and the digital people should work more closely with the TV people (instead of constantly slagging them off).

    Insights can be gained from the type of analysis you showed in your presentation and insights are valuable. However, the greatest value is derived by uncovering what you can not explain, because that creates the opportunity to learn. (See Consultants and Confused Apes)

    – Greg

  11. October 12, 2009

    This is getting tedious, but here we go:

    1. I am not doing predictive modeling, no. I am doing process modeling, which is very different.

    I’ve already explained why I don’t believe that predictive modeling is cost-effective (or effective at all). Read my previous comments again if you missed it the first… and second time.

    2. Communications planning does fall into the model, yes. It is a component of it. A component is not the whole. I think it’s pretty clear that I am not just talking about communications planning. Not sure how you missed that as well.

    3. Greg, if you don’t think I define R.O.I. in my presentation, I have to question either your sanity, your I.Q. or your attention span.

    But hey, I’ll give you the benefit of the doubt: Try slide #16 If the equation isn’t clear enough, try slide #15 for a more visual explanation. If those two slides still don’t clearly define R.O.I. for you, I don’t know what to tell you.

    4. You want to look at net sales first, then market share. As an aside, I never said that competitive activity doesn’t affect sales numbers or market share. I’m not sure where you got that from.

    Look, it’s clear that your expertise is in billboard marketing, and I respect that. You probably know more about billboards than I ever will, and that’s cool and all, but I don’t think you’re ready for this level of discussion. I feel like I am saying one thing and you’re hearing another.

    If you at least countered with arguments that made sense, this would be interesting. But you are clearly either not understanding my comments, or not taking the time to actually read them. Either way, you don’t seem equipped to be a participant in this debate.

    5. As for the final part of your argument, I agree with your comment that “having everybody running around trying to calculate precisely their contribution would just create an enormous mess.”

    I think you should know that I neither suggested it nor implied it. Again.

    So look, this is my last comment to you on this post. Feel free to have the last word if that rocks your world. I won’t mind. If at some point in the future, you really want to learn how to do this stuff, you know where to reach me. But first, you are going to have to do a few things:

    1. Pay closer attention to what people actually say, not what you think you hear them say. If you don’t completely understand something I write, read it again. Don’t hastily fill-in the blanks with what you THINK I wrote. Likewise, if something seems to be missing from a presentation, go through it again to make sure you didn’t miss it all on your own. You have misunderstood more about my presentation and comments than you understood, even though most of it was excruciatingly clear. I’m not sure why that is and I don’t really care to speculate. All I can tell you is, whatever the problem may be, slow down and pay closer attention before misinterpreting other people’s work, statements, etc.

    2. Sometimes, it’s okay to admit that a topic is beyond our level. I’ve been determining R.O.I. for projects and programs for 15 years, long before social media existed. I know this stuff inside and out. I’m not an agency-side guy, I’m a client-side guy, which means I know the ins and outs of cost accounting and P&L management. Among many other things to help me learn how to do this exceedingly well, I’ve managed a $120M business for a $7.5B corporation. I know this stuff like nobody’s business.

    On the other hand, I don’t know jack diddly about billboards, and I have no problem admitting that.

    So perhaps, just perhaps, you might want to consider the possibility, however minuscule it may be in your universe, that my understanding of ROI (and measurement as a whole) as it applies to business (not just media) is vastly more advanced than yours? Maybe? Just maybe?

    I am trying VERY hard not to be mean, but Greg, I think that when it comes to this topic, you are completely out of your element.

    3. You need to let go of your preconceived notions and false assumptions about a number of things. They’re holding you back. If you don’t know what I’m talking about, go back to my previous comments, and this time, read them very slowly so you won’t miss anything.

  12. October 12, 2009


    At this point, I think it’s best that we agree to disagree.

    Best of luck in your business and your life!

    – Greg

  13. October 12, 2009

    This was not just a blog post, it was a commitment!

    I liked this discussion a lot. From an outside observer, you two are not that far apart. You both emphasize accountability in a field that is often not being held accountable, and that is the most important point.

    I agree that Olivier came across as a bit heavy-handed in this presentation but I have been reading his stuff over a period of time and if you look at the overall contribution of his work, in my opinion (coming from academia and 27 years in Fortune 100 sales and marketing) — he gets it. In a short presentation, it’s OK to be less thorough and more entertaining in order to make a lasting point, so you have to consider the whole picture.

    Another thing to consider (and where I think you two also agree) is that even non-financial indicators of success (awareness, readership, downloads, page views, etc.) still lead back to an attempt to measure financial impact. It may be difficult and elusive to pin down ROI but that should not keep you from trying.

    Jamie Wallace made an analogy in a blog post ( that I think is wonderful — social media measurement is like being a great bartender. Sure, sales are important, but there are so many other intangibles that contribute to success and you can’t ignore those, either.

    And here’s the good news — there is so much to measure! I have been spending time learning about the latest developments in social media measurement and in my mind, this is the most exciting area of marketing right now. You guys will have a lot to spar about in the future.

    Greg, I’m glad you introduced yourself to me. Your perspective is fresh and important. Keep putting it out there.

  14. October 12, 2009


    I don’t have anything against social media. I just think that communication planners should do their jobs and social media people should do theirs. That’s why big brands hire give companies like ZenithOptimedia communication planning contracts:-)

    – Greg

  15. Stuart Nicholson permalink
    October 12, 2009

    I have found this debate stimulating.

    I actually think that the fact that people are prepared to commit time and column inches to such passionate defence of their positions proves that this issue is one of the most important that communications professionals should address.

    I am sympathetic to elements of both of the views here and do concur, as did the previous contributor, that the two of you are actually not far apart.

    I just have three observations:

    1.Econometric modelling can isolate contributing elements within communication and prove extremely valuable in focusing emphasis on the parts that work.However, as has been said, it often isnt feasible on the grounds of absolute cost.In my experience, even among large companies, the number that have invested in a suitained exercise of this kind have been very few.

    2.ROI isnt just about measuring results.Its also about focusing properly on the objective.With the exception of direct-reponse based clients it is all too common to find that communications plans are briefed without a clear focus on this and leads to a brief to try to do too many things, some of which are less critical in their contribution to the desired result.All too many clients do this.Often an “ROI” in this instance is whatever metric after the event looks to be positive whether or not this was what was actually intended or needed.

    Companies whose business model leads to a communications solution that is mainly digital, particularly in search, usually know exactly what they want to achieve and measure success precisely.Those that still have a high proportion of traditional media in their mix have more difficulty in deriving this focus.

    3.You cant easily isolate the performance of digital unless its your only communication vehicle.Just becuase the “last click” comes from a digital platform, doesnt mean you can attribute 100% of the effect to that element, as other forms of communication have often contributed to the final consumer action.I do tend to come down on Greg’s side in this part.Becuase the digital part is measurable and the non-digital is less so, it leads to a tendency to attribute a disproportionate effect to the digital platform in delivering the ROI.Thats not to say that i have the answer here, as this separation is difficult to arrive at,but we do need to recognise this.Without giving too much away, one way is actually to measure the consumers experience itself and quantify the contribution of each element that way.

    Anyway…keep up the debate.Its interesting to follow.

  16. October 12, 2009


    Thanks. Coming from the “ROI Agency,” your opinion means a lot.

    I agree 100% that ROI needs to be clearly defined. If it’s limited to response, it is less problematic (although response online can be affected by offline media. Heavy TV campaigns are well known to affect CTR’s).

    However if, as is usually the case for large clients, there is a large integrated marketing effort there needs to be an integrated ROI effort. Once you have Digital Media people measuring the effect of offline media they cease to be Digital Media people and start being Communication Planners.

    Communication Planning, as you well know, is a full time job. If all of the Digital People become communication planners, than who does Digital Media?

    – Greg

  17. October 24, 2009

    A truly fascinating exchange of words and intelligence between two or more “experts” who clearly know their stuff and have strong opinions about it. Gratitude for being privy to it all.

    As one who is fascinated by ever-changing technology and how it impacts basic communication and relations between human beings (for we are the only species that cares about the internet) and one who has been involved in enrollment activities and business development for over nine years… I have a question for all of the participants of this multi-layered dialogue.

    And I am honestly deeply curious on your personal philosophies on best practices for getting to know new people and peers:

    At what point does one resort to old fashioned technology like a phone call or a meet up and take an online conversation offline? Or as Olivier suggested, meeting at an event being attended by both? Conferences are great catalyzers for making new friends and colleagues clearly.

    Posts and comments on blogs fascinate me for they allow others (like myself) to be flies on the wall or be looking through the two way mirror (and remain undetected till as reader we also comment.)

    And in doing business development of any kind, use of digital technology for communication can both delay or prompt closure and commitment. It’s a duality and a paradox.

    And the internet has created opportunties for people to meet people from near and far, but how well do we really know someone if all we ever do is email or comment on blog posts? Language is both the cause of problems and the solution to it.

    Virtual intimacy and authentic relating: can it be cultivated digitally and deeply if people never take it to the next level of ? Would this debate have carried on longer or ended differently had it happened face to face? I’m curious.

    A veteran of publishing and advertising turned business development specialist with a passion for lifelong education, self-development, and authentic marketing, I am deeply curious about what ultimately convinces and converts a curious prospect to commit and make the purchases they make; be they small or large investments.

    What turns an acquaintance met in a Yahoo Group or on a blog into a future “Friend” in the ludite sense of the word.

    I heard it said very elegantly by Marty Marsh, a business coach, that we must remember behind EVERY transaction that goes on on the internet there is always a human being. And the internet is a tool for advancing human relations and connections. And yes, it’s people who buy things.

    How many rounds of comments does one go before saying: “care to talk in real time?”

    To cut to the chase and lively debate, now that conversation I would happily be a fly on the wall between this dialogue in real time.

    And if I have wandered off topic please forgive me. It’s simply because whenever I read or experience sales and business discussed in terms of numbers, data, ROI and so on, I cringe a little.

    For if the makers of products and services forget even for a moment, that it is people with feelings and problems that need solutions who are buying products and engaging in services, they will lose the potential business of the person who doesn’t want to be treated like a bit of data and a number.

    A fascinating analysis and discussion to be privy to all the same!

  18. October 24, 2009


    I think you’ve touched on an important topic that explains the rise of communications planning over the past few decade. It used to be that the message was primary and Creatives reigned supreme.

    There was no point in doing much communication strategy when You could reach half of the country with an ad on the Ed Sullivan show. The only opportunity like that that still remains is major sporting events (which explains the enormous resource that goes into Super Bowl ads).

    When cable TV gained traction in the 80’s, audiences started to fragment rapidly and by the 90’s communication planning started to take on greater importance. In effect, you had to start taking into account who was listening before you started talking. Digital technology has taken this even further.

    So, I think the important thing is to be listener focused. In effect, there is no point in a telephone if the line is dead.

    The issue is related to ROI because the “line needs to be open” for sales to be stimulated. While direct response is important, people filter out most brand messages. There are simply too many of them to pay attention to. The stronger the brand, the higher the response will be.

    I came across a good example when I was involved in some planning for McDonald’s. The operations people always wanted to do sales promotions. However, as brand attributes in the tracking data fell, sales promotions became less effective.

    In effect, at some point sales promotions were stealing from future sales. When they did brand promotions, they got worse sales results for that month but better sales for future sales promotions.

    I hope I answered your question. Let me know if I didn’t.

    – Greg

  19. October 24, 2009


    Appreciate the thoughtful reply and response to my addition your dialogue! I do have one last question for you, which I will pose at the end of this follow up comment.

    A few things about me that might put my ludite comments into context.

    I gave up watching television in 1989 in the traditional sense (or cable programming for that matter) more than once and since 2004 have again been television free. The internet provides access to all the news and entertainment I want and need! I can seek out what I want for myself, and choose to read a lot, research for fun and subscribe to a whole lot of ezines and blogs. Add in some radio for color, crunch and texture. And printed periodicals are near and dear to me still. And I enjoy selectively: film and theater and live events. But I am clearly discerning about what input I choose. I get this about myself.

    And so, I manage to stay abreast of many (but not all) current world events mundane and otherwise within this powerful, transparent and digital age. I am first and foremost an avid reader and researcher. I am also the sort that likes to engage with brands and people I feel warm to and appreciated by. I am a tough customer who values impeccable customer service. And I treat clients like I like to be treated.

    Ironically, part of my past professional years include print production and managing direct mail, designed to attract customers and foster relationships. As many as possible from a targeted mailing in a single communication effort. So I get what you are saying about the rise of TV and advertising.

    The printing press is one of my favorite inventions! Once upon a time it was the highest form of technology for communication no? And word of mouth has been an effective strategy for centuries as well, yes? The internet accelerates communication.

    And these days my best work is done communicating one to one or to smaller groups. Mass communications rarely get my reaction. But then I am a deeply relational person. And I want to believe I am not a dying breed; that there are generations living and to be born that will be interested in authentic relating to their immediate environment and community as well as communities at large that they identify with.

    And the most fascinating thing about advances in technology related to communication, mass communication and social networking sites and applications etc. is how it “seems” we are all getting better connected and acquainted.

    Are we?
    But there are a lot of “conversations intended to lead to more” that simply fade and dissolve.
    Sales that don’t happen.
    Good ideas that don’t thrive.
    Relationships of all kinds that don’t stick.
    And no transaction of any further kind happens.

    I realize you are addressing the concerns of business in your blog and the science of it. Data can tell a story.

    I was fascinated by the volley of comments between you and Olivier Blanchard and here is my question:

    At what point do you suggest to him or me or anyone that “phone lines are open, let’s connect” and talk in real time vs keeping the conversation in a series of chat comments?

    When does connecting one to one become the next best and important thing to do in business in an age where mass communication is easy for us all to engage in?

    Am I being completely naive or asking a very large philosophical question?

    Thanks! And pardon my random abstract thinking, for this too is how I apparently am (smile).


  20. October 24, 2009


    I think it is a very good question, and one that everybody needs to answer for themselves. The overarching media trend is towards greater diversity in communication. For example, while you say that you don’t watch TV, viewership is at all time highs as is magazine readership.

    As Mark Penn pointed out in his book Microtrends, there it only takes 1% of the population to create a viable market, even a hit. One of the great things about the communication revolution is that we can communicate any way we want and, increasingly, we are able to find people who feel the same way.

    So I would say that if you would like to communicate with someone in a more meaningful way, just ask them if they would like to. They can always say no.

    – Greg

  21. October 24, 2009


    I do that very thing all the time. And you are right they are at choice as to saying yes, let’s or no thank you.

    The Long Tail by Chris Anderson is also a great read.

    I will continue to follow you on Digital Tonto.

    And when compelled, comment.

    Thanks for engaging “Digitally” in what feels like a real conversation that is for the moment complete.

    Happy writing!

  22. October 26, 2009

    Deborah I love this:

    “At what point do you suggest to him or me or anyone that “phone lines are open, let’s connect” and talk in real time vs keeping the conversation in a series of chat comments?

    When does connecting one to one become the next best and important thing to do in business in an age where mass communication is easy for us all to engage in?”

    I also appreciate Olivier’s comments and of course how Greg responded to all this barrage of ideas and opinions on the ROI and social media.

    Let me just say a few things and very short.

    1) ROI is not what we need, we need to find out what operators affect our bottom line…. and here SM is part of the marketing funnel….

    2) Based on 1 above, things will differ for B2B versus B2C – and, culture matters – no way anybody will sell printing services in the UK on weekends using Facebook, print shops are closed ….

    3) Whatever you do, establish a baseline of your social media activities and then watch the trends and compare to what others accomplish.

    4) Before we do all the above, let us establish about three objectives what we need to accomplish or better, what kind of problem will social media help to solve and then benchmark …. did we do it or did we fail.

    For me running a B2B outfit we did the above and SM is working but the company has to sit down and figure out much on its own…. what is it what we want to get out of air conditioning, office chairs, company cars, fancy receptions and/or sponsoring Polo events for our clients.

    Nevertheless, if I cannot see that social media helps us as well as some of our clients improve customer engagement and satisfaction, out it goes…..

    So far so good ….

    Thank you for an inspiring blog post and great comments


  23. October 26, 2009


    Good points. Thanks for your input.

    – Greg

  24. October 26, 2009

    Enjoy the common thread addressing ROI is not feasible. I would be willing to bet with each disagreement, each approach to ROI is totally different.

  25. October 26, 2009


    I agree, but I would point out that I never said that ROI wasn’t feasible at all, just not feasible for specific actions that make up a minority of total activity.

    Generally, marketing ROI is not impossible and, in most cases, valuable. However it does need to be well defined.

    – Greg

  26. Ibrahim Moss permalink
    October 31, 2009

    Ok you geniuses, tell me about iPods and iTunes.

    (lets not forget that part of iPod’s popularity is the iTune interface, and the content selection contained therein.)

    Tell me which one markets the other. Tell me what metrics we use to determine the ROI of one when compared against the other.

    (If I remember correctly, Apple outsourced the build-out for iTunes. The original contract cost was a little more than $600,00. I suspect service / maintenance costs are a fraction of initial build outs. Apple does not directly market iTunes, but this UI generates well over a billion in revenues a year (gross unadjusted). Apple does market its iPod & iPhone lines. I do not know the costs.) Its revenues for iPod sales are about 2.6 billion / quarter.

    Gets complicated right?

    Ibrahim Moss

  27. October 31, 2009


    You should be more specific – Launch year or subsequent years?


  28. Ibrahim Moss permalink
    October 31, 2009

    I mention the iPod / iTune example because I really believe it has, whether we know it or not, become the norm for a new kind of ROI analysis. In short, I believe we may have to develop new indicators and measurements to determine any ROI for products and service offerings that are introduced in the digital sphere. For me, it is just too complicated and wholly inappropriate to use traditional matrixes to calculate ROI’s in the digital space.

    But, to answer your question, the answer is a bit more complicated than it seems, and I do not have all the numbers. iTunes launched one quarter behind iPod. So, their respective first year revenues include about three quarters sales for iPod and two quarters of sales for iTunes. Really, that part of the calculation is simple. One simply adjusts the numbers for part year sales.

    As for subsequent years, I am not sure of the costs for maintaining the iTune server and back-end costs, research, IU design work, licensing department issues, etc. If I were to gander, I would say that these internal costs for iTunes are at 15% of revenues. (I am not going to mention revenues actually realized from actual iTune downloads. I believe Apple acquires a royalty for these iTune sales. These royalties are likely less than 25% at best, of actual sales revenues)

    The bigger question is kind of a chicken v. the egg thing. Which product is really driving revenue? Is it that iPods drive revenue for iTunes, or is it the opposite? The answer is a little more complicated than it seems.

    (Take into consideration other efforts by the rest of the competitors. Zune, for example, has matching products, but its music store is not well developed. Amazon has a well developed music store, but no matching MP3 player. Pure off-line word of mouth promoting, social fad , cool factor, recognized white / gray ear buds, etc (this is a kind of advertising that is not different than wearing a Nike tee shirt), all contribute. On the other hand, the easily navigable UI on the iTune site, along with content, are extraordinary drivers of sales.)

    There are other factors as well. What may be asked of the iPod / iTune phenomena, may be applied to the iPhone, or other Apple products. For that matter, one may have to really factor the increased value of the Apple brand in driving all its sales. That brand value is exceedingly important, and likely exceedingly impossible to calculate.

    Having said all of that, I believe when the smoke clears, we suddenly see that all the stuff about ROI’s and other indexes are wrongly applied the digital sphere. We simply need new formula to calculate returns. We need formula that reflect values individuals and place on any given product or service, as well as his or her effectiveness in communicating the same to a social network when making a recommendation to purchase a product. After that is done, we can measure the efficacy of any business’ marketing effort, ad product, etc. (I know that can be complicated as well. It varies from video to print, to banner, to repeats) 😉

  29. October 31, 2009


    I agree. ROI is an enormously complicated issue for an integrated marketing program. There are a variety of approaches, and none are perfect.

    – Greg

  30. Ibrahim Moss permalink
    October 31, 2009


    I reread your first thread in this ROI blog. I think you get it.

    But there is a wild world out there that is tracking differently, and with meritorious intentions. Look at Avenue A / Razorfish. Well, there business model is, in part, founded on determining the metrics of measuring digital ads, and thereafter using such to determine advertisers’ market strategies. Razorfish’s initial approach mated, in part, users’ demographics to ad products Well, that service is highly valued by very smart people, even senior ad executives who have been in the business for decades. Is Avenue A / Razorfish’s contribution of no value? (btw, they just sold. This is a highly valued company).

    There are other things out there that raise some interesting questions in this digital ROI era. Take Google for example. It now places ads based on, inter alia, relevance to the content of one’s e-mail message. This approach may significantly affect one’s ROI for digital marketing. It is distinguishable from Razorfish’s approach in that its placement engine mates with e-mail content rather than the e-mailer’s demographics.

    There is a lot of stuff to be sorted out for sure. Its all art now; the science may catch up later.

  31. October 31, 2009


    I think a distinction needs to be made between response and ROI. Digital Media is very good at measuring response, but that can be affected by many things outside of the campaign as well.

    Moreover, having a good response campaign won’t help you if your brands internal attributes are falling. That’s a different problem and requires a different solution.

    – Greg

  32. November 15, 2009

    There are so many wholes to this argument I cannot list – this is what ad execs said in the 1950s. Only thing I can say is that if Google followed this dusted off “digital market is the same” hypothesis, the company would no longer be in existence today.

  33. November 15, 2009


    Thank you for your input.

    – Greg

  34. Pan Aveyard permalink
    February 11, 2010

    Hi Greg

    I’m inspired by the passion involved around this subject, the above goes to prove that.
    The ROI of the social web/media will be danced around for some time, with various methods put into practice.
    The one thing that surface throughout though is the need for a baseline. Without having the intelligence first, gathered through optimised searches of all platforms, any talk of measurable return is pointless.
    Having systems in place to access all this is where the strength lies.

    Intelligence is King..


  35. February 11, 2010


    Good point.


    – Greg

Leave a Reply

Note: You can use basic XHTML in your comments. Your email address will never be published.

Subscribe to this comment feed via RSS