Why You Should NOT Follow Procter’s Latest Marketing Advice

Marketing is transforming from top down to bottom up and you need to embrace it, not run away.



Editor’s Note: Marketing ROI, mix modelling, attribution, programmatic, cross platform measurement and similar topics are hugely important debates taking place right now. In fact, it’s so important that we’ve partnered with Sequent Partners, Time Inc, the MMA, and many other brands, agencies, tech platforms and media networks to develop a new event: the IIeX Attribution Accelerator forum.

As marketers, we all want to understand what touchpoints drive consumer purchase behavior – and allocate our marketing efforts and resources accordingly. Marketing measurement techniques like marketing mix modeling (MMM) and digital attribution enable us to do just that, but both approaches need to evolve to show us the full picture.

Mix modeling needs to become more granular, timely and actionable – like attribution. Attribution needs to be more comprehensive, addressing the entire marketing mix, and needs more scientific rigor – like marketing mix modeling. As attribution moves beyond digital, and marketing mix modeling moves beyond traditional, a more integrated approach to marketing measurement is needed.

In today’s post, Joel Rubinson (who will be presenting some ground breaking research on this topic at the AAF event) tackles this issue head on, using P&G’s recent shift back to traditional marketing as an example of how brands lack a clear and confident vision of how an integrated digital model can deliver for them. It’s good stuff.


By Joel Rubinson

Recently, Procter & Gamble sent shock waves throughout the marketing community when it announced it was abandoning precision targeting via Facebook.  “It didn’t work”, “We targeted too much and went too narrow”, said Marc Pritchard, their CMO. All generalized from a Febreze ad targeted to pet owners in large families that didn’t seem to cover its CPM costs.


Of course, when this was reported in the Wall Street Journal, and picked up by the marketing press on earth and neighboring planets, it was big news. It was interpreted as generalizable advice that all CMOs must listen to…do not target too tightly!! Do not move too far away from the old way of marketing. Some even blogged they hoped this will bring down ad-tech.

My advice? Don’t let P&G set your media strategy. Marketing is transforming from top down to bottom up and you need to embrace it, not run away.

Data-driven precision targeting is a train steaming down the tracks. Programmatic is reported to be growing at 20-50% per year. Mobile too…right place right time.  Programmatic is purely about delivering the right message to the right user at the right time, decided on in real time…something Google refers to as “micro-moments”, or Clayton Christensen might refer to as “jobs to be done”.

Programmatic is transforming marketing from “top down” (buy the whole audience of a show you feel matches your brand) to “bottom up” (choosing to buy an impression to a given user based on their profile and situational factors.) Yes, bottom up marketing is transformational; for example, hypothetically, Hillary Clinton can now runs ads to undecided voters on Foxnews.com, avoiding hardcore Republicans…before programmatic, who would have thought that? This is how Presidential marketing now works, the most sophisticated target marketing on the planet.

Yes, there are growing pains…some issues of brand fit with the media property, transparency on fees, whether CPMs are priced right, and the occasional counter-intuitive result like Procter got with Febreze, etc.

Push past the growing pains and double down on bottom up marketing. Most of my consulting is really about helping marketers or research companies develop the insights and analytics side of what this means. (Yes, insights and analytics is MORE important than ever in a data-driven marketing world.)

Doubling down on bottom up marketing (and research) opens up wonderful new possibilities:

  1. Target the 15% or consumers who are likely to account for 80% of the trial of your new product, taking money from the 70% who have virtually no interest in your brand whatsoever.
  2. Target persuadables towards your brand and not waste ad dollars against those firmly committed to other brands
  3. Activate your consumer segmentation by making it the basis for constructing audiences you target programmatically that deliver higher ad response (and if that doesn’t happen it means your segmentation approach is in need of a digital mindset refresh).
  4. Turn your DMP into your brand tracker. If you have modeled propensity scores on the DMP, you actually have a brand equity metric at scale that can be reported out continuously and in real time. Add in social media and you have a surrogate for attribute ratings (You should be able to save millions of dollars if you have numerous brands in numerous markets).
  5. Response profile your brand.  Instead of just attribute/positioning profiling, let’s profile out the distinguishing digital profile variables that differentiate those who respond to your advertising. Then turn that knowledge into audience creation rules so that you continuously improve the response to your marketing investments.
  6. Treat your campaign as a living lab with real time decision making. As Procter found out with Febreze, marketing doesn’t always follow script. Oh those wacky, lovable consumers! Every ad campaign can be optimized by moving money around in flight based on what is working, whether it conforms to preconceived notions or not…and is an opportunity to learn something…don’t squander that.

And one more point…when Pritchard said, “we targeted too narrow”, in particular resist that advice. The future of marketing IS narrow…it’s bottom up marketing, one impression at a time, perfectly chosen. It doesn’t mean that you are restricting reach…it just means you are building reach from the bottom up. In fact, I predict that even the big shows and publishers will all unpack their audiences and develop targeting plans for you that are bottom up.  It is already happening. And the reach you need is not necessarily the reach that Procter needs, so again, build your own experience and media strategy.

It’s becoming a bottom up marketing world…understand it, embrace it, master it, measure it.

You can leave a response, or trackback from your own site.
Tags: , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,

13 Responses to “Why You Should NOT Follow Procter’s Latest Marketing Advice”

  1. Chris Robinson says:

    September 6th, 2016 at 7:07 am

    Seems there has been problems with correctly identifying the audiences in these precision targeting systems. https://www.facebook.com/salut.ro/videos/1208049929217082/

  2. Ellen Woods says:

    September 6th, 2016 at 9:57 am

    When I read this, it seems like there is a comparative between apples and oranges. Aside of that issue, there are issues around assumptions and context that are driven by the need to target market within a mass marketing objective.

    Are we talking a success of impressions or of conversion? Obviously, it’s conversions but the benchmarks are largely driven by impressions targeted in a multi-media exposure in both the P&G issue and the article both seem to refer to the success of the advertising. THAT is the problem.

    It’s about the message translating. The average exposure is less than nine seconds to any content. That means a heavy reliability on the visual. It also means that consumers have to have the issue, not have a solution or a poor solution and are planning a shopping trip where the product is available while they are playing around on Facebook.

    That’s a lot of what-ifs.

    It is too narrow for CPG. It is also too narrow for most commodities because they don’t stay top of mind. When you had print ads, people would tear them out and put them with shopping lists or in a coupon box. There was a physical response that solidified the visual and neuro response. Digital advertising is like being on a Farris wheel with the colors of the world whizzing by.

    Joel is right that bottom up targeted marketing is needed because there is no adequate way to measure mass marketing in a world full of vehicles. However, P&G is also right in that a narrow focus makes a lot of assumptions that don’t prove out.

    While there are a lot of ways to approach targeted marketing, it should be noted that CPG depends on mass acceptance and products must meet certain thresholds to hold their spot in the production line. They compete with outside products and also with internal rankings.

    It’s hard to target market when your target customer necessarily needs to represent a wide variety of conditional variables, thus the too narrow comment. What does seem clear is that new products or variations on old products need to be promoted at the store level with the digital content pushed closer to the ZMOT.

    I am not sure what type of campaign they were running but it seems like a digital coupon program would be more effective in the short term. Perhaps a step down with higher values at the start. I think you have to have a physical connection in CPG to drive that connection and it has to be broad enough to quantify. Then you can target. It’s not like TV where the whole focus is on the ad and unless they interact there is no way to know if it resonated. There are just too many campaigns running.

  3. Mark George says:

    September 6th, 2016 at 9:02 pm

    It seems to me that you’re saying that we should all embrace bottom up / targeted marketing because….1) we can and 2) lots of people are doing it. Am I missing something, but I don’t see anything in here on its effectiveness.

  4. Chris Robinson says:

    September 6th, 2016 at 10:26 pm

    Ellen, you raise some serious issues and I would add to those.. Many. ad displays using video have a 3 seconds before you can skip. For may viewers its an annoyance and the eye fixation is on the second countdown. The problem is this would be counted as an impression because it only requires one second of exposure to be treated as an impression by the industry audit boards. Can I repeat that – one second! Now given lack of attention and annoyance one has to question just how valuable that impression is. I think we know the answer, not much at all. Lets add to that ad blocking and known levels of fraudulent viewership claims (25% on Chrome alone!) and you start to get a sense of how little value we can attach to any viewership measurements.

    The industry response has been to ignore all this and build some internal validity around data by using programmatic models, which are essentially just gilding the crap that is reported social media data. Any wonder companies like P&G are not getting results. One only has to look at the experience of other brands. A year ago Kraft General Foods reported their belief that over 75% of claimed ad impressions just never reached a human or were just fraudulent. Like any industry it is in no one’s interest to blow the cover on this or the money disappears. Would anyone expect an ad agency or media buyer to challenge the data? As an industry we should be highlighting this factual data. The WSJ and NYT have been all over this. Its in eh public domain for god’s sake! Even WPP are wondering if the diversion to social may have been misplaced. If they are wondering, it must be patently obvious.

  5. Chris Robinson says:

    September 6th, 2016 at 10:32 pm

    One other thing. That Google ZMOT is a load of rubbish. No one is reporting actual data based on observation. It’s all based on smartphone user claims about researching prior to shopping (no frequency, no product links) and claims of doing price comparisons in shop. We get the same misleading data in China, but I can tell you, you rarely see a smartphone in the hands of any supermarket shopper. Its just rationalized, expected answer, “aren’t I smart” rubbish. Until someone can prove with real usage data, I think we can ignore this as another social media myth.

  6. Joel rubinson says:

    September 7th, 2016 at 9:45 am

    Thank you for your comments and adding credibility to the importance of the debate. To be blunt, I feel like traditional researchers are all on the spaceship in Hitchhikers guide to the galaxy, reinforcing your sense of importance among yourselves without regard for the outside world. I live in two worlds…marketing research and digital and I can tell you, the conversations are totally different with marketing leads with digital marketing responsibilities or with data scientists in the multi-touch attribution space. start talking to your media agencies. Start thinking about the fact that programmatic is now 30% of many ad budgets and soon will be half or more. Few if any marketers have measured how much reach can be achieved digitally and I bet they would be surprised how broad that is..,probably rivaling TV at this point. Do you know how big mobile is? I bet not..because Facebook and google advertising are more than half mobile but they do not break that out. All Facebook and Google advertising is precision targeted. Final point..dear researchers…try to learn something from passively occurring digital data. Your insights don’t have to all come from a survey! The world I am describing is not just precision targeting it is also a marketing research data point…millions if not billions of them. It’s a new marketing world… Embrace it…get off the rocket ship.

  7. Mark George says:

    September 7th, 2016 at 10:14 am

    Joel, thanks for ‘playing the man not the ball’. My point is simple. And there is no vested interest here. You give us plenty of reasons why you feel we needWR8B to embrace targeted bottom up campaigns. But you do not talk about effectiveness. ‘Everybody’s doing it so should you’ isn’t reason enough.

  8. Joel rubinson says:

    September 7th, 2016 at 1:17 pm

    mark, I am advocating that marketers get serious about evaluating a new way of marketing from the bottom up. That means try it using a test and learn mentality. I am optimistic that it will lead to an improvement in marketing ROI form all the evidence I have seen working with clients on proprietary assignments but don’t take my word for it, try it and measure it. The irony in your comment is that ironically, digital is the MOST measurable of all media because ads can be tagged and matched to users. You can then use multi-touch attribution or RCT to measure the return. TV and other analog media offer nothing close in terms of measurement precision.

  9. Joel rubinson says:

    September 8th, 2016 at 8:10 am

    And for those who still don’t get that the future is bottom up marketing, herer is the latest news from Target. They are doing what WalMart did. they are getting into the ad tech business by using their frequent shopper data as a DMP, allowing them to offer precision, bottom up targeting to their product vendors like Unilever or Coke. This is not only a great precision targeting capability but it will transform shopper marketing and category management. The future is here and IS becoming widely distributed! http://adage.com/article/cmo-strategy/advertise-a-target-shopper-luck/305753/

  10. Chris Robinson says:

    September 12th, 2016 at 8:37 am

    Joel instead of assuming we market researchers seem disconnected from marketing reality why don’t you address my claims that the numbers are misleading and frankly fictitious. I think this is highly relevant to a chief marketing officer, but no one is telling him this or even raising concerns. Big difference between Walmart data and that from social media giants like Google and Facebook who may well be perpetuating impressions as some kind of indignation of real viewership. At least offline media has validity and reliability plastered all over it. You seem to be too focused on the story telling. Start looking at the data – or isn’t that what you do? Funny because that is what we do when riding in our spaceships.

  11. Larry Gibson says:

    September 14th, 2016 at 11:53 am

    In the end, the critical customers, those who make a difference, are the customers who are at the cusp of choice, the customers you have lost narrowly but could win plus also those you have won narrowly but could lose.
    This critical ‘segment’ can only be identified one customer-at-a-time. It’s size and composition are a function of the specific advertising effort. It can only be very grossly identified by any external characteristics.

  12. Ellen Woods says:

    September 20th, 2016 at 11:09 am

    Joel, I appreciate that you live in two worlds (as do most of us in the marketing world) and that budgets are being directed more toward digital (primarily because it is the only vehicle that can be measured at this point) but measuring reach in a dynamic environment is pretty analogous to measuring how far you can swim in the ocean. While it may be better than most people can do, it isn’t very far in terms of your distance toward a foreign shore.

    The problem with targeted digital marketing is that it isn’t an exclusive club. Anybody can and many companies are doing it and everybody is trying to one-up everybody else. I went to Target, Costco and my local gas station yesterday and they all pinged me to let me know they knew I was there. Offers? Yes. Irritation: Huge. Google pings me to remind me to pay bills I have already paid. So, I turn off push notifications and they invade me from a different point. If I had trouble reminding myself to buy dog food I would create a shopping list with Alexa. I know turn off my phone when I go in a store and I don’t browse anymore, I mission shop, in part because I have other things on my phone and don’t want it off for too long.

    This is a long winded way of saying that targeted marketing works pretty much like a sniper shoots. It should work in a way that benefits me. That means they have to actually interact with me, not my profile. What we have today is targeted mass marketing, not personalized marketing and while that is not surprising, it has the opposite impact of what they are trying to achieve.

  13. Tim Burke says:

    September 21st, 2016 at 8:36 pm

    Joel – thanks for this. Ever since I read this article in wsj I had similar conclusions. Niche targeting is the future. Given that Proctor had success by broadening its targeting simply tells me that they hadn’t properly identified ALL the different niches that their campaign would have resonated with. The lesson here isn’t to give up on hyper-targeting, rather highlights the importance of truly knowing your audience with high fidelity at the beginning of any campaign.

Leave a Reply