eBay’s search engine marketing team used to be famous (infamous?) for what can only be described as SEM carpet-bombing.
In the early days of Google AdWords, it seemed like eBay showed up on every term a user entered, regardless of whether it actually had anything to do with products available for sale on the auction site. We’re talking terms like “nuclear bomb” and “belly button lint” – not exactly the definition of “commercial intent.”
Over time, Google has restricted advertisers from uploading the dictionary and then using dynamic keyword insertion to make an ad look like it’s relevant to a user, largely through the enforcement of Quality Score, which penalizes advertisers for irrelevant ads and keywords.
The only minor modification to this policy came about a couple of years ago when Google introduced remarketing lists for search ads (RLSA), which theoretically allows an advertiser to bid on an irrelevant term if the advertiser knows that the person searching for that term has visited the advertiser’s site recently. So with RLSA, eBay might again decide to buy the word “belly button lint” if it knows that the searcher recently visited the auction site to research personal hygiene products.
RLSA is Google’s first foray into offering a behavioral overlay on top of search; it encourages advertisers to expand their horizons from just “intent” to a combination of intent and audience.
If Google were to let advertisers leverage first-party data in search it could be one of its most lucrative advertising programs ever.
Facebook’s Approach
Just a few miles down the road from Google, Facebook has taken a much more aggressive approach to behavioral advertising.
A couple of years ago, it rolled out “Custom Audiences,” a feature set that enables advertisers to upload first-party customer data – things like email lists, addresses, and phone numbers – and merge it with Facebook’s customer data. Advertisers could then market to their actual customers on Facebook, as well as to “lookalike” users that Facebook’s algorithm determined were similar to the advertiser’s best customers.
And last month, Facebook announced a partnership with DataSift to allow advertisers to run ads based on the content of users’ actual posts. So an advertiser who wants to reach people who are, for example, writing posts about “belly button lint,” now has the ability to do so.
In other words, Facebook advertisers now have the ability to run ads based on a combination of:
Their own first-party data, via Custom Audiences as well as cookie-based first-party data;
Facebook’s psychographic and demographic user data (second-party data);
Third-party data (via Facebook’s partnerships with companies like DataLogix);
Content in users’ posts (via DataSift).
By comparison, an advertiser on Google AdWords search can use:
Cookie-based first-party data (RLSA);
Intent derived from the query a user enters.
If Google wanted to, it could offer targeting on search that would come close to rivaling Facebook’s offering. For example, Google already offers psychographic and demographic targeting on the Google Display Network, so it’s reasonable to assume that this could be ported over to AdWords.
Google also offers plenty of behavioral targeting functionality on GDN, similar to Facebook’s DataLogix partnership, and Google could accept first-party data feeds from advertisers to help them better target customers and potential consumers on SEM.
To date, however, Google has not released any such products. To my knowledge, no one at Google has publicly spoken about why it hasn’t done so, but murmurs I’ve heard suggest a heated debate is going on internally about the right line to draw between user privacy and advertising revenue. This is no doubt a legitimate debate (though I’ve personally long abandoned the notion that I have any privacy at all when I’m online).
Big Money In First-Party Data
My prediction is that the advocates inside Google who are pushing for greater use of first-party data in the name of higher advertising revenue will eventually win this argument. And when they do, it has the potential to be one of the most lucrative advertising programs Google has ever launched.
Consider a term like “belly button lint,” with little to no commercial intent. Other than the RLSA scenario outlined above, the odds that an advertiser would ever show up on this query are slim. So every day, millions of searches without commercial intent result in zero revenue to Google.
Imagine what would happen if Google allowed advertisers to upload first-party data like email addresses and postal addresses (and then have this matched to Google user IDs). And imagine further that Google created a lookalike algorithm to help advertisers find new customers who closely resembled their best consumers.
In this world, the term “belly button lint” would likely be filled with advertisers, not because they want to sell a belly-button cleaning solution, but because they know that the person who typed in this query is a great customer for whatever they’re selling. The ad shown to the user would be as diverse (and query-agnostic) as the ads currently shown to someone scrolling Facebook’s News Feed.
Monetizing The Unmonetizable
The opportunity to monetize currently unmonetizable queries is probably the single biggest untapped advertising opportunity online today.
Allowing advertisers to leverage first-party data would also increase competition on queries with commercial intent, for two reasons: First, because advertisers would bid more for existing queries if they knew that the searcher was an actual or likely customer.
And second, because many advertisers would begin to bid on commercial intent queries that aren’t actually relevant to their product, but are being searched by their target audience.
So those eBay ads that were simultaneously ubiquitous and irrelevant 10 years ago? You might start seeing them again. But they’ll be highly relevant and likely have nothing to do with belly button lint – unless of course, that’s your thing!
Comments