skip to main content

“On the Internet, nobody knows you’re a dog” – Still True Today?

Created at April 14th, 2014

“On the Internet, nobody knows you’re a dog” – Still True Today?

It’s hard to believe this iconic cartoon came out over 20 years ago.  For those of you still mastering the alphabet or busy at your weekly AYSO games in 1993, “On the Internet, nobody knows you’re a dog” by Peter Steiner was published in The New Yorker on July 5, 1993. As of 2011, the panel was the most reproduced cartoon from The New Yorker.

on the internet nobody knows

20 plus years later Peter Steiner’s cartoon still resonates, in particular for those of us in the digital advertising space. Do advertisers today feel confident they can recognize and reach their customers in the digital world, where consumers are spending more and more of their time and money with brands of choice? The technology to identify customers in digital has gotten exponentially better since 1993 with the advent of the cookie, first supported by Internet Explorer (then the ubiquitous browser) in 1995. But cookies fall very short, they are strained and accuracy is increasingly called into question.  If I’m browsing on 3 devices and 2 browsers, I’m still 1 person, but in the cookie world I’m 5. It’s a problem! Also, there are growing reach implications for the advertisers as the new releases of many major browsers are blocking the setting of 3rd party cookies by default. The question top of mind lately for advertisers when trying to determine the efficacy of their advertising campaigns in the digital world, and other channels is match rates.

For companies delivering customer segments to advertisers it’s become a dog fight over match rates.  The term “matching” is often defined as the percentage of overlap between an advertiser’s customer file and a cookie pool. A he says/she says over match rates have driven the advertisers to be focused on the wrong question. Before asking a partner what their match rate is they must define how they are matching their customer IDs. Is it with a premium publishing partner who knows their registered audience or an intermediary technology platform fundamentally architected on 3rd party cookies? The match rates can vary significantly with these different baselines. The next question should be on the accuracy of the match, which gets to how a partner does their matching. Is it via email address or string based only? Or is it predicated on something more robust like linking, e.g. the technical process to “match” based upon a direct relationship with a premium publisher. Is it individual level or household level? With string base matching, you do a direct match so, for example, if one record shows “William” and “Bill” on the other, it won’t match.  If you’re matching on email only, think about how many email addresses you have today and how they are used.  It is much better to do matching where we can identify that William Smith at one address is the same Bill F Smith at another address.  And it’s best to do individual level matching and recognition so you can then aggregate to household level as appropriate.  But you’re in control of those decisions.  Finally, an advertiser should consider their partner reach, as scale is paramount. What is the scope of individuals and/or households a partner can match to the advertiser’s desired audience? A match rate is meaningless if it’s not measured against a high degree of accuracy, and the rate becomes irrelevant if you’re not delivering it at scale.

These initial questions will help clear up a lot of confusion for advertisers on match rates, and the process for matching in different channels. It’s a safe assumption you’re going to have a significantly higher match rate in a direct channel where you are targeting a customer based on a publisher’s knowledge of their registered users vs. targeting a cookie.  In the digital channels it’s imperative for an advertiser to know not simply their partner’s match rate, but their process for recognizing their customer in this channel, often called recognition, and defined as the ability to identify a unique individual through at set of possible signals such as email, cookie ID, postal address, log-in or handle. You can identify a large addressable audience on all the premium publishers, but the only meaningful metric is how many of those you can recognize as part of your audience of desired customers. Also, ask if your customer is being recognized at the individual vs. the household level, to determine the level of accuracy, and if it qualifies as 1:1 marketing. A partner’s ability to identify and reach those desired customers with a high degree of accuracy and scale is what you really want to know, not their non-specified match rate.

A dog may appreciate anonymity online, and protecting that is critical to earning a seat at the table. But even the dog still wants the relevance and context from his favorite brands, and for this you do have to sell to an individual (or in this scenario a dog), not a cookie.