Billions of dollars of wealth have been generated in the valley in the last small number of years based on the largely unquestioned premise that targeted social media ads work. Such targeted ads are very arguably both Facebook’s and Google’s secret sauce – not to mention that of the large ecosystem of companies around them.
But is it actually true? Are these ads truly a kazillion dollar a year secret sauce that merit the premiums that brands pay for them? Not according to the world’s largest consumer brands company and global advertising spender P&G (Procter & Gamble.) P&G is known for being fastidious measurers, for being both a good process company and an innovative one. P&G may not be a household name, and may not be especially sexy, but it is virtually impossible to live in the Western World and not have at least several of their products in your home. P&G has a large number of brands, everything from shampoos to toothpaste to diapers, soap to toilet paper, razors to cosmetics.
In yesterday’s WSJ, there was a front page article noting that P&G was generally discontinuing the use of targeted ads on Facebook. The below is a short excerpt from it:
“Procter & Gamble Co., the biggest advertising spender in the world, will move away from ads on Facebook that target specific consumers, concluding that the practice has limited effectiveness.
Facebook Inc. has spent years developing its ability to zero in on consumers based on demographics, shopping habits and life milestones. P&G, the maker of myriad household goods including Tide and Pampers, initially jumped at the opportunity to market directly to subsets of shoppers, from teenage shavers to first-time homeowners….
P&G’s shift highlights the limits of such targeting for big brands, one of the cornerstones of Facebook’s ad business. The social network is able to command higher prices for its targeted marketing; the narrower the targeting the more expensive the ad.”
In other words, P&G found that targeted ads didn’t help it sell toothpaste and other daily essentials. Objectively, there is little reason to think that it would. Facebook may well not know – or be able to infer – what toothpaste you use for example. Nor is there any real reason to believe that you are likelier to purchase a given brand of toothpaste because your friend uses it.
To be fair, P&G isn’t lowering their overall spend with Facebook – at least not for now. And, as the WSJ article duly notes, some specific, highly directed types of targeting (e.g. expectant mothers for diapers) are effective. But if the targeting in the general case is no longer perceived as an advantage for which large brands will readily pay a hefty premium, it isn’t clear that Facebook will be able to command either the percentage of advertising spending that they do now. Likewise for Google.
The main thing that is interesting in all of this is how long the assertion that all of this just magically works has been effectively unchallenged, even in the face of common sense and day-to-day experience. When is the last time that you deliberately clicked on a product ad? Weeks? Months? Actually bought something that originated with such a click?
Sure, it is true that seeing an ad for a given brand somewhat increases the chance that you’ll buy it – but only if it is a category of thing that you anyway buy, and you don’t already have a strong preference for a competing brand. But this raises the interesting question of whether the combination of a proliferation of ads, and a generation that spends a good part of its waking hours glued to one screen or another, will result in the latter adapting to the former to such a degree that most of these ads effectively become invisible to them. It is highly doubtful that anyone is researching such a possibility, since it would be hard to imagine who would fund such a thing.
For all the moaning and groaning about big data and the invasion of privacy, the reality is that most ad targeting that works is of the simple “if single-thing-X then…” variety such as the expectant mothers example above. Much of the rest really doesn’t work very well, even with the massive amount of data being collected and the billions of dollars spent to analyze it. There are three main reasons for this:
Lack or sparsity of data: Everyone may buy toothpaste for example, but few people talk or tweet about their toothpaste. And the reality is that many more things are at the toothpaste end of the spectrum than the fashion one. For many types of products, that just may not be a fixable problem.
Limitations of current technology: To understand how limited the current technology is in many respects, just consider how many of the ads you see are for things that you have already long since bought. For example, when my coffee maker broke, I bought a new one pretty much immediately. Most people would do the same. Yet for literally months afterwards I was barraged with coffee maker ads. This might make sense for a luxury purchase that many buyers may ponder for months, but not a commonplace small appliance that many people consider a necessity. Or a few months ago, when I was looking for a power of attorney form for a particular state, for weeks I got useless ads for attorneys. I am now getting ads for cemeteries – I truly have no idea why. I could go on and on. In short, it is still mostly the low hanging fruit that is being picked, even billions of dollars later.
Business Incentives: Obviously Facebook, Google et al are strongly incented to sell what are tantamount to “qualified leads” to their advertisers. The more factors that are considered in the targeting, the fewer targets there will be for which to charge a premium. More targeted ads are more expensive, but that’s irrelevant if the target pool to offer for sale is miniscule. While of course the advertisers have some say over what dimensions and parameters to target, they are limited by tools, by the available underlying technology, and doubtless in many cases by expertise.
None of this is to say that FB, Google, etc aren’t highly valuable advertising platforms even despite this at this point, just that their advantage over other forms of media and platforms is objectively less than what had been claimed. For example, having comparatively less data about their users for ad targeting purposes was one of the big knocks against Twitter. I doubt that this overdue revelation will harm the stock price of either FB or Google much, and certainly not to the $20B extent that one of the large pharma companies was punished in market cap for the clinical trial failure of one of its drugs recently. For one thing, there is not as direct a relationship to revenue, and FB will likely be quick to push customer testimonials that rebut P&G’s findings. And we are talking advertising, not blood tests: it may not actually matter whether the much touted thing works or not.