If Ad Targeting Works, Is More Ad Targeting Always Better?

A former colleague at Forrester Research published a study that described ad targeting and his view of the future of online marketing.

The study is available to Forrester clients only, but the crux of the argument is laid out in a blog post: Why Google – Not Facebook – Will Build The Database Of Affinity.

The author, Nate Elliot, describes large media companies building a “database of affinity,” which he described thusly:

Recently we described an idea called the database of affinity: A catalogue of people’s tastes and preferences collected by observing their social behaviors on sites like Facebook and Twitter. Why are we so excited about this idea? Because if Facebook or Twitter or some other company can effectively harness the data from all the likes and shares and votes and reviews they record, they could bring untold rigor, discipline, and success to brand advertising.

Nate’s argument as to why the database of affinity represents the future of online marketing is this: online search was revolutionary for marketers because it allowed them to target their customers better than ever before. Marketers could now spend advertising dollars towards ad targeting, getting their brand in front of people actively searching for their products (e.g., Samsung advertising to searchers of [buy LCD TV]) rather than passive watchers of TV or print.

In keeping with the thinking that more targeted is always better, Nate argues that brands will be able to use historical affinity behavior of consumers to target more precisely than ever before.

It’s clear that the database of affinity is a prize of gigantic proportion, and ultimately, Nate concludes that Google will end up the winner in that battle because they have the broadest data to draw upon and best chance of making sense of it.

I want to set aside the question of who will win that battle and question some of the assumptions upon which the affinity database argument is built.

Is More Ad Targeting Always Better?

Nate’s argument rests on the assumption that if some ad targeting is good, more ad targeting is better. Thinking about this in isolation, at the level of “theory,” it stands up to scrutiny. But there are places where, at the level of “practice,” it seriously falls down — and in my view, these flaws are sufficient to make me question the validity of the whole argument.

For example, not long ago, I bought a pair of jeans online from Bonobos. A little while later, I was shown an ad for Bonobos in my Facebook feed. I was perplexed by the timing of the ad targeting, until I read an article that described brands’ practice of providing Facebook with customer email addresses that were then mapped to Facebook logins so that targeted ads could be shown to brand customers. This is technically more targeted, but I responded negatively since I had never opted into their feed.

I think the question this raises is this: when it comes to ad targeting, is there a point of diminishing returns? Maybe there is such a thing as just enough ad targeting and no more.

Are All Players Equally Able To Leverage Affinity Information?

Any discussion of which media powerhouse (Google, Facebook, Twitter, etc.) will win the affinity database race presupposes that all players have the same ability to act upon that information. But, context matters.

In search, I am in research mode: in entering a query into the search box, I am announcing my intent to accept “bids” for my attention from all takers, and I’m prepared to act upon the provider of the best information in the search results to that query.

In social, I am in a very different state of mind. I expect content from my closed network, and unless I have invited a brand in by Liking or following them, their presence is very much viewed as an uninvited intrusion. It’s the loud, slightly drunk guy at the party that keeps interrupting your conversation with the interesting guy/gal.

It’s possible that users will become more accepting of ads appearing in social streams and begin to specifically turn to social networks for a wide variety of information retrieval scenarios — for shopping/product research, job search, etc. This would give brands a strong opportunity to leverage the database of affinity in marketing on social networks.

But as the chart below shows, this is not yet occurring at significant rates. This all adds up to our remaining highly skeptical that purely social companies have the ability to effectively leverage any database of affinity they may be able to build.

information retrieval social vs search

The Real Question Is “What Is The Context?”

Finally, I want to question a concluding assumption Nate makes in how the affinity database must be leveraged if marketers are to be successful:

If marketers are going to use affinity data to power brand advertising, simple text-based ad units won’t cut it. Brand advertising demands large, video-based ads to create discovery — TV spots, pre-roll ads in online videos, and supersized online banners.

I don’t buy the argument that large, video-based ads are a minimum requirement to reach consumers. Context matters — there is more than one way to reach an audience, and Google’s $29 billion search advertising business shows that text-based ad units do, in fact have their place.

Database Of Affinity: Great In Theory, Maybe Less So In Practice

The idea that digital marketing is progressing to targeting via a database of affinity makes sense. The argument had me nodding my head in agreement initially, but further reflection produces some serious doubts about the assumptions upon which it is built.

First, there is some question about whether more targeted is always better; and second, there is a question of whether all players involved have the ability to effectively leverage a database of affinity.

What do you think?

A version of this article originally appeared in Search Engine Land

5 Responses to If Ad Targeting Works, Is More Ad Targeting Always Better?

  1. Adam Dince says:

    Great article Nathan. Here are a few thoughts:

    1. I’ve been using highly targeted ads on Facebook and it’s allowed me to build solid communities for engagement. I think the level of success a brand has depends on what it’s offering. Selling plumbing supplies on Facebook probably won’t provide much ROI. However, selling affinity type items like apparel, music, etc… has a great opportunity to succeed. I’ve read about a few entrepreneurs who have been making magic happen with the right kinds of social advertising (and I’m sure there’s more than just the few I’ve read about).

    2. Sentiment is changing. When social ads started to infiltrate Facebook, they annoyed the hell out of me. Now, I find myself more open to them (mostly because they’re more tailored to my interest). Additionally, I’m more used to ads being a part of my social experience, so I don’t mind them as much and am more inclined to click.

    3. Paid promotion of content on Facebook is a great way to drive traffic to your Website. Once you get the traffic to your Site, it’s up to you to make it pay off. I can tell you that I volunteer for a non-profit and I drive a ton of traffic to the Site from Facebook for a nominal cost. From those visits, we’ve got tons of leads, sign-ups and other important conversions we need.

    I think there is a movement to demonize paid advertising (and vice-versa on the paid side), which I think is foolish. Paid advertising works. Paid and organic work even better together. Sure, organic/earned is ideal because it costs a lot less. However, often times, you need to pay to get the reach for organic/earned to work at the scale you need to make an impact. Especially as many networks including Google and Facebook give more real-estate to paid advertisements.

    Keep up the great work my friend.

    • Nathan Safran says:

      Thanks for the always insightful comments, Adam. I agree that there is a time and place for everything, but also that those who might be too reliant on paid media to feed traffic and conversions might rethink how to start capturing some traffic organically.

      Be well!

      • Adam Dince says:

        Agreed. That’s been a problem for years. The smart companies/brands/agencies are figuring out the right mix. On a side-note, when brands hire people to manage their earned and paid media practices, that are not qualified to do so, things get really out of hand. And I see that happening a lot these days.

  2. Mengyu says:

    Great article!

    “It’s possible that users will become more accepting of ads appearing in social streams and begin to specifically turn to social networks for a wide variety of information retrieval scenarios — for shopping/product research, job search, etc. This would give brands a strong opportunity to leverage the database of affinity in marketing on social networks.”

    I think this will be the future trend. I feel the targeting ad on Facebook now is too simple. The advertiser didn’t use their time and effort to attract their targets. What they do is just like the example you gave us. You bought a pair of jeans online from Bonobos and you saw the ads of the Bonobos jean again on Facebook. This is annoying and people will sometime feel they are being interrupted.

    However, if the advertisers could use the “database of affinity”, they know what style of jeans this person likes. They could send them a jeans research or an article about how people would like when they wear the jeans, for example, his/her favorite celebrity wear the jeans on the street. This might be a incentive for them to buy the jeans and they might like the ad and not feel they are being interrupted.

    I also like the concept “diminishing returns” of the targeting ads. It will be interesting is advertiser could know that. Some people who really good at research could start collecting data to see if there is a diminishing return point. I am curious about whether people could use number to measure their effort spending on ad targeting.

    Well, thanks again for the great article!