Marketing Mix Modeling (MMM) Follow-up

12 Jan

Last June I wrote on P&G’s work with Delahaye in marketing mix modeling (Original Post).  There have been a lot of interesting comments (see comments), and I wanted to take a moment to answer a few questions posed by ph hanky in his/her January 10 comment.  Here is the comment:

ph hanky – January 10, 2007[Edit]

who does P&G’s MMM work anyway? Delahaye isn’t a modeling agency.

The problem with companies like P&G and Delahaye making sweeping statements like this is that there is no one to audit or verify the findings. Anytime you have someone doing a regression analysis like MMM, you have to question if the regression analysis was done with the right vigor and make sure enough factors were factored in the analysis.

P&G has been making lots and lots of money and AG Laffley has been talking about their MMM and how they are cutting back on TV in favor of other vehicles. I don’t doubt P&G pulls off some world-class MMM, but I don’t know if they promote stuff like this to put the screws on their ad agencies and TV advertising rates.

If they spent half of what they spend on TV on PR, then they would saturate the airwaves, the web, and print with PR. They don’t address diminishing returns on PR. I would assume you reach saturation point with PR long before you reach saturation on Tide commercials.

First, the easy question.  The company that worked with Delahaye to develop the model is a firm called Communications Consulting Worldwide ( 

Regarding the audit or verification of findings, my personal view is this is not the right way to think about modeling.  A properly built model should yield a highly correlated approximation of the relationship between public relations results and downstream outcomes, usually sales.  There are certainly some issues that the modeling company may find uncomfortable – co-dependence of variables and reverse correlation to name two.  So while there may be a little leap-of-faith with the model, the rigor is generally much greater than is normally applied with simple media content analysis – and the leap of faith to ROI is much shorter. 

The other issue ph hanky raises is diminishing returns on high levels of PR investment.  In my experience we should be so lucky.  I have never seen a completed model that demonstrated PR investments were on the diminishing returns side of the curve.  On the contrary, every model that I have seen suggests PR investments are well below diminishing returns.  If a company is spending $60 million in advertising and say $1 million in PR (actual numbers from a former client), then the issue of diminishing returns on PR is not applicable.  Moving just $1 million from advertising to PR would show nice leverage in ROI.  I would love to be in a situation where I had to go to a client and recommend we reduce PR spend because we have reached diminishing returns, although I not holding my breath on this one. 

Thanks for reading, DB

9 Responses to “Marketing Mix Modeling (MMM) Follow-up”

  1. Peter Verrengia April 6, 2007 at 5:12 pm #

    Just seeing the reply to the hanky post now, but I’d like to offer a correction. Communications Consulting Worldwide did not create the markeing mix model for P&G. P&G did it in house, working on four brands, and I think Delahaye provided data for at least one of those brands. If you want to see an example of what CCW does do, check out the Feb 12 PR Week case history on how Southwest Airlines measures the business impact of communications. And as for the modleing techniques, it is not entirely clear to us what P&G did in its marketing mix model, but in the work we do at CCW we use causal modeling and we find it is very possible to demonstrate the impact of communications activities on sales, earnings, stock price, or any other business outcome without the problems you describe. Peter Verrengia

  2. Don Bartholomew April 6, 2007 at 7:06 pm #

    Hi Peter,

    Thanks for your comment. This issue was already corrected with this post (

    Thanks very much for reading. -Don B

  3. ahndunk February 16, 2008 at 1:07 pm #

    Good post, All information here are newsworthy.

  4. MJW May 27, 2008 at 5:10 pm #

    The company that does P&G’s MMM is Analytics Partners. As far as I know, that is their only client.

  5. Sachin Kapoor May 30, 2008 at 8:14 am #

    Marketing Mix Modeling is a statistical analysis technique used by marketers to understand the individual and combined contributions that multi-media marketing investments have on business results. Thus, marketing mix models help estimate the ROI associated with historical marketing spending, as well as forecast the prospective business results that future spending will generate.

    Marketing Mix Modeling involves breaking up of sales volume into various components, and analyzing spend on each of them to calculate ROI from each of these components. After knowing ROI at different levels of marketing activity, threshold and saturation levels, one can forecast sales through each of these activities and hence optimize the marketing spends to gain maximum value.

    In the last 10 years many CPG companies have adopted MMM. Many Fortune 500 companies such as P&G, Kraft, Coca-Cola and Pepsi have made MMM an integral part of their marketing planning. This has also been made possible due to the availability of specialist firms such as Fractal Analytics, who have developed Marketing Mix Modeling solutions to help their clients optimize their marketing spends and hence improve their brands across markets. Statistics show that with same budget, clients have improved their brand sales by up to 35%.

  6. John Dawson May 30, 2008 at 3:03 pm #

    One point i’d like to add as a practitioner in the dark art of “MMM” – I’ve looked at the impact of PR within models a few times for major multinationals and one thing not reported here is the difference between positive and negative PR. No data available but suffice to say, in both cases the “lost value” from negative PR had an enormous impact on sales performance – one far greater than the positive upside from the good stuff. The key takeaway for me was the need for the organisation to act quickly.

    One other point to make, I’ve always found it hard to equate positive PR coverage with spend on PR – far harder than with media where you get what you spend. For this reason, I’ve yet to see anyone put up the PR curve and say “you need to spend on PR to get to here” – it doesn’t appear to work that way.

    One other point – the data I’ve seen clearly shows diminishing returns existing with PR as with almost all channels.

  7. December 26, 2008 at 12:05 pm #

    Sir or madam, if you continue to follow this blog, what is source for your info on Analytic Partners? Is it the same firm here: I am interviewing with them now and I have understood they have other clients as I am also seeking a client side position in a firm which has made mention of utilizing them during interviews. I would like to have any information that may help me make career decision.
    Thank you,

  8. rana zaki October 17, 2010 at 4:52 am #

    can any one please , know how to measure the 4s of the web marketing mix, if any one can, please tell me

  9. Joan November 3, 2011 at 2:35 am #

    Gracies per compartir aquesta informació. Es molt interessant. Et convido a que vegis els meus darrers treballs sobre publicitat a

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: