User identity graphs have a critical role to play in digital advertising in the near future. Vinod Kashyap, head of product, Digital Element, shares why marketers need to use user identity graphs for improved branding and marketing.
If you’re a marketer, you’ve probably heard about user identity graphs (aka user ID graphs) and the critical role they are predicted to play in digital advertising in the near future. Over the next 12 months, assuming Google doesn’t delay the deprecation of third-party cookies again, savvy marketers will test user ID graphs against their third-party data to see which delivers the best results.
There is a reason user ID graphs are receiving a lot of attention: they’re a privacy-safe alternative to proxied third-party data. Many believe they’ll ultimately deliver stronger campaign performance than cookie-based behavioral audience segments ever could.Â
What exactly is a user ID graph? It’s a database that associates all of the “identities†of a user to a unified ID. Your campaign may see “me†on my work computer, home laptop, and mobile phone. By stitching together my unique identities (e.g., email, mobile ad identifier, browser cookie) into one unified ID, you can recognize me whenever I encounter your brand in the digital universe. From a marketing perspective, it also means if you have my email address, a user ID graph will allow you to target me on my mobile device.
See More: How to Tame Cloud Identity Sprawl
User ID graphs also link offline records, such as physical addresses of your customers in your CRM data, to online identities. This is helpful if you want to target, say, your past customers for a reactivation campaign, as some ID graphs can associate home or work addresses to online devices. In this scenario, you would hash all your customers’ addresses and send them to your user ID provider to match against its data. You can then target against all matches.Â
Interestingly, this approach is based on your first-party data, which means it’s less proxied (read: more accurate). It’s also more privacy-compliant as you’ve presumably requested permission to use customer data for marketing purposes. Â
Who Are the ID Graph Providers? Why Should Buyers Beware?
User ID graphs hold tremendous promise, which is good news for digital advertisers. But like all promising new technologies, many companies want to get in on it, and therein lies the risk. Not all user ID graphs are created equally. Complicating matters further, advertisers and marketers don’t have much experience in assessing user ID graphs and the user data they offer. This means they may conduct tests, get disappointing results, and write off the approach prematurely.
What do you need to know to make smart decisions? Many of the user ID graph providers are companies that have a history of seeing a great deal of Internet traffic, e.g., LiveRamp, The Trade Desk and other DSPs. These providers tend to have massive datasets and can promise high match rates (i.e., the provider can match a high proportion of your users to their IDs, enabling you to target them in other channels and devices).Â
But many other providers have less visibility into the Internet and must purchase their data from other companies, which can be problematic for many reasons.Â
Firstly, does the provider have enough data in the region where you want to target? The provider may have a lot of data on U.S. consumers, but that’s not helpful if you want to reach people in Canada.Â
Second, where that data is collected from will determine its usefulness for your campaign. Let’s say a provider collects mobile ad IDs (MAIDs) from a large swath of cooking apps. That provider may have millions of MAIDs in its graph, but if you want to target auto intenders, that graph won’t depend on your ideal audience.
Finally, beware of “match rate†promises. High match rates are often touted as a reason to use a specific provider, but there are multiple ways to look at it. Media agencies prefer to work with companies that offer the highest possible match rates because their goal is often to reach as many people as possible. If you’re launching a brand awareness campaign, a high match rate is ideal.
See More: Identity Mapping Traps and How to Avoid Them
If you want to home in on specific users in the hopes they’ll convert or take a specific action, a high match rate can be a disadvantage. A high match rate may mean you’re paying to target many people who aren’t suitable for your product and will never convert. Thus, your cost per conversion will be higher. This, by the way, is precisely why you should ask providers how they source their data. A lower match rate may not be bad if the provider’s data matches your campaign’s regional or audience criteria and is therefore likely to have a higher conversion rate and lower cost-per-actions.
Even if you want to launch a brand awareness campaign, the graph with the highest match rate can still be problematic if its attributions are weak. This is a complex concept, so I’ll give you an example. Let’s say I used my laptop to purchase an item from your website, and as part of that transaction, you’ve captured my IP address. Now let’s say you want to target me for an ad on my mobile device, and you ask a provider to find me based on the IP address you captured. Sounds great, right? But what if I made that purchase while I was at a Starbucks in a city or state far from where I live? Your IP address is Starbucks’, not the one assigned to my home WiFi. I might never go back to that particular Starbucks again. The graph may have a match between my MAID and that Starbucks’ IP address, but how useful is it?Â
The upshot: It’s easy to purchase bid-stream data and use that to populate a user ID graph. But ask yourself: how is that data attributed to a specific user, and how strong is that attribution?
Questions to Ask Before You Buy
Here are some questions you can ask a user ID graph provider to see if their product is right for your campaigns:
- How do you source your data? Is it transactional data? Geographical-based data, such as IP addresses? Are you getting it from cookies or other digital footprints?
- What regions does your data cover? Can it meet my campaign’s criteria?
- Is your data deterministic or probabilistic?
- How do I know you’re attributing this data to the correct or relevant customers?
- What elements are included in your ID graph? If, for instance, you want to reach your customers or prospects as they stream media, your IP address is an important element to have in the ID graph you use.
It’s a good idea to test multiple ID graphs to know which works best for your brand, desired channel, and campaign goals well before the third-party cookie finally goes away. Not all ID graphs are created equal, nor do they offer the same elements and channel strengths. Fortunately, you have the opportunity to get ahead of the game.
Are you tailoring your digital marketing strategy to user ID graphs? Tell us your thoughts on FacebookOpens a new window , TwitterOpens a new window , and LinkedInOpens a new window . We’d love to know!