Two Changes That Mean Visitor Analysis is Dead

Unidentified Flying Visitors
I'm alarmed by the evolution of an old issue.

There has never been a way to create a "360° view" of an individual's interactions with a brand. Businesses have mostly decided that they have to live with that gap in understanding.

This topic has been addressed many times, and there are many reasons for it. However, we are now at a tipping point that requires serious reexamination.

Big Changes

There are two huge problems now that have never had quite the same impact in the past: people use multiple devices and people use multiple channels. Fundamentally, both of these problems have always existed, but their impact has been very limited.

Now things have changed.

A Plethora of Devices

A plethora of devices
People now use multiple devices to interact with companies. Smartphone sales account for at least half of all phone sales in America now. Tablet sales have skyrocketed since the iPad was released. Media PCs (eg. Google TV, Apple TV, Roku, Boxee) have risen in popularity in the last year. And people still use laptops and desktops.

This explosion in device proliferation has happened all at once and almost entirely in the last year.

Multiple Channels

The lines have blurred between online and offline behavior. With the expansion of social media and mobile apps, people can interact with a brand without the brand moderating the interaction or even knowing it is happening. People form opinions about a company without the company ever explicitly doing anything. People could always get their friends' advice in the past, but this is a scale and magnitude never seen before—yet it's impact is almost equally untraceable.

Why You Should Care

The unprecedented scope and scale of these changes has silently morphed the definitions of even the most basic metrics. The problem is, we do not know what all the new definitions are. There are no signals in our reports to tell us how much it is skewing our data right now. Ironically, this comes at the same time that we have more data points available to us than ever before. The result is that we blithely continue to interpret the reports using outdated definitions.

A few examples of how this impacts visitor-level data:
  • Attribution models get thrown out the window. Visitors use multiple devices now to interact with the website. Every device is seen as a new visitor. Especially where sales cycles (or the equivalent) take more than a single visit, we now have no framework for what it looks like.
  • No measure of social media's impact. We can see shares and clicks on shared links, but that is all. How this relates to a person's overall relationship with a brand can't be tracked.
  • Number of visitors is unreliable by an unknown degree. This includes visitors to the site, visitors in a segment, visitors exposed to a test, etc. How can we ensure we have statistically significant findings? We just do not know how much any particular segment is being over-represented or what the impact of that is on our analysis. 
  • More difficult to understand a person's needs and intentions. We lose all context. We have very limited insight into how to optimize an experience when we can't connect what all of their interactions with the brand were.
  • Visit-to-visit personalization is unreliable. Without knowing what they did last time, we cannot recommend what they should do this time.

Even visit-level analysis is not immune to these changes. Some examples:
  • Not clear what type of device is being used. The lines between phones, tablets, netbooks and laptops are blurry, and they aren't getting any clearer. It is even more difficult to differentiate a media PC from a laptop or desktop.
  • Even if we know the device, it might not mean anything. Device type won't tell us what situation the device is being used in (eg. with a group, during a brief wait, on a long flight, etc.), and every situation may demand a different experience. The possibilities have multiplied.
  • Website visit does not happen in isolation. The visit may complement unattributed things that a company is doing.
  • Companies need to interact with people, not instances. Companies cannot focus on visits. They need to understand how to meet people's needs rather than optimize an experience for a predefined use case.

Some Solutions

Solutions to the problem of fuzzy visitors
There are a few solutions that occur to me, but none of them are perfect. I think finding a solution is possible, and I'm anxious to get thoughtful ideas from others.

For example, forced authentication would solve a lot of these problems. With the right incentives, a person would always want to be logged in when interacting with a company's site or app. This would give a more complete picture. It is not a solution for all companies, but it is a necessary solution for some, even though it doesn't answer all of the concerns.

Companies can survey customers to get a better understanding of how people interact with a brand. At best, though, this would only give estimates, and they would be aggregated. This could definitely give actionable insights for a company, but it wouldn't solve for the biggest problems. Also, because the landscape is shifting rapidly, it would need to be repeated frequently.

Perhaps the most realistic solution is to leverage a third-party. For example, Facebook and Google Plus integrations provide incentive for most users to be logged in. If it were possible to capture a unique identifier for each visitor, it would allow companies to see a more complete picture of how visitors interact with their brand regardless of which device they used. One potential downside is it would require deeply integrating third-party platforms into the entire brand experience.

Is this becoming an issue in your organization? Are there other negative impacts you are seeing? Do you envision other solutions to this issue? Are these concerns completely over-hyped?

No comments:

Post a Comment