Measuring unique people with web analytics…
by Matt Belkin
posted on 05-11-2006
In my past few posts, I’ve suggested that Unique Visitors is a flawed metric, prone to significant inaccuracies. I realize this position is controversial. In fact, one blogger suggested that it was downright “irresponsible”. That’s fine – there was a time when everyone thought the earth was flat. It takes time, often a great deal of time, to accept that the traditional view may have been wrong. It’s also fascinating to me that despite their pushback, these naysayers have yet to offer any reason why Unique Visitors should be perserved as an accurate metric. Anyone? Buehler?
Ironically, even though I’ve heard the most pushback from vendors who regrettably base much of their analytics on Unique Visitors, these inaccuracies are not vendor-specific; rather, they are largely manifestations of the Internet itself as I highlighted in my recent post, 15 Reasons why all Unique Visitors are not created equal.
I’ve also suggested that due to these limitations, I often recommend that web analytics professionals use Visits or Sessions in their baseline analysis, as it provides a more accurate and dependable view of conversion, persuasiveness, and ultimately Return on Investment. And to clarify some recent feedback I’ve received, this includes both macro and micro conversions.
For example, if you operate a high-touch retail site where many visits often occur before purchase, you should look to not only macro site conversion (orders/visits), but also micro conversions such as product configurator conversion (success product configurations/visits). For jewelery sites this configurator could be a “ring selector”; for Travel sites it could be a “saved itinerary”, for Automotive sites it could be a “vehicle configuration” or “dealer quote”, etc.
All of these represent micro-conversions; which are effectively subsets of your macro site conversion and represent key milestones or decision points in the customer lifecycle.
In turn, I’ve acknowledged that Visits also has its own limitations when it comes to marketing optimization initiatives like clustering and segmentation, remarketing, latent response, lifetime value, campaign ROI, and other “person-centric” analyses.
Due to these limitations with both Unique Visitors and Visits, I’ve hinted at the concept of Unique Customers. In my opinion, Unique Customers addresses many of the significant inaccuracies inherent to Unique Visitor measurement, but also compensates for the lack of uniqueness when using Visits as a metric. So what is a unique customer and how do you measure one? Read on…
What is a Unique Customer exactly?
Unique Customers can be derived using several different approaches – but almost always relies on an individual identifier like user “123456” or “mbelkin”. These identifers are typically assigned to your visitors during a purchase, registration, subscription, lead submission, or support inquiry process.
For example, let’s assume you run a retail website called Electronics Plus. When new customers purchase from Electronics Plus, they are asked to create a new profile with their personal information. This information, while not always required, typically aids in customer support and order execution, among other benefits. When the customer has completed their personal profile, Electronics Plus assigns them a unique identifier like “mbelkin”. This login is usually secured with a password so the customer can protect their purchase history, order status, or expedite purchases in the future. At the same time, the website itself typically encrypts this unique identifier so it cannot be easily comprised – for example, “mbelkin” would become customer “17894653”.
So let’s consider this for a moment. This simple checkout process has just yielded an individual ID that in most circumstances, can be used to identify a unique customer. And importantly it does not suffer from the same inaccuracies of Unique Visitors. In most cases, the ID will not change if visitors delete cookies; it will not change if multiple visitors use the same computer; it will not change if one visitor uses multiple computers; it will not change if visitors reject cookies; it will not be impacted by IP pooling nor will it be impacted by dedicated corporate IP addresses.
This individual ID also does not suffer from the same limitations of Visits; namely, the lack of uniqueness. Rather, each Unique Customer is truly unique – connected to one person, irrespective of time, place, or technology.
In short, this approach is among the best you’ll find for identifying truly unique people and prospects that visit your website. Furthermore, this methodology lends itself to much more robust marketing analysis, customer segmentation, and lifetime value analysis because it is connected to an individual. And perhaps most importantly, since remarketing to individuals is a primary strategy after you’ve completed marketing analysis or customer segmentation, this unique identifier is critical to executing those campaigns.
What if I’m not a retailer?
Many of you do not actually sell products and services online – and that’s OK. Unique Customers is not a retail-specific concept. Media publishers can leverage this strategy thru subscriptions and/or member registrations. Lead generation companies in the automotive, finance, telecom, and technology verticals can leverage lead forms, customer logins, member registrations, product registrations, or any number of other events. Support websites can also leverage the registration process to establish this identifier, and most already do today.
A word on privacy
Wait, what about…?!
At this point, you probably have a myriad of questions about unique customers. Should this metric replace unique visitors? Should it supercede it whenever available? What about the millions of visitors that I can’t collect a unique ID for? What about micro-conversions? What about behaviors that occurred before the visitor identified themself? How does this impact my KPIs? What new KPIs should I look at? What new optimization possibilities does this offer?
These, and many others, are all good questions with good answers. If you’d like to discuss this topic further, please do not hesitate to contact me and the Omniture Best Practices Group.