Prospected Traffic Quality Report

What is the purpose of this report?

The Prospected Traffic Quality Report is a tool for marketers to evaluate how effective their prospecting channels are at delivering new users who are likely to purchase. Platform metrics can be unreliable, sales cycles can be long and require touchpoints with multiple channels, sales metrics can be volatile. The report attempts to simplify your analysis by reporting only on a new user’s first session, and how your model graded that first session. It’s a proxy for what you really care about, which of course is purchases, but it’s a reliable proxy from which you get immediate feedback, and which you can use to objectively compare different channels to one another. More on how the model grades a new user’s first session below.

What do the metrics in the report mean?

There are two key measures underpinning the report: prospected users and promising users.

Prospected users

A prospected user is a new user to the site who was acquired by the reported traffic source. We define a user as "new" based on the user identifiers we track, which might differ a bit from Google Analytics or other similar platforms.

Since we only evaluate a prospected user’s first session, this metric is synonymous with what we might call prospected sessions; that is, first sessions from new users. Any subsequent sessions that a new user goes on to have will not be taken into account.

Promising users

A promising user is a prospected user whom the model graded as being reasonably likely to make a purchase in the future. Since your model is tuned specifically to your site, what your model defines as “reasonably likely” will be custom to your site, and may change over time. The majority of your future purchases will come from promising users, even though they tend to make up a modest proportion of total prospected users.

Promising user rate

The promising user rate is simply the count of promising users divided by the count of prospected users. Again, this is a metric that you can easily use to compare different prospecting channels to one another.

How do your clients use this report?

The most natural use case is to compare prospecting channels to one another. How does your promising user rate on Facebook compare to Instagram Shopping? Considering the cost of acquiring new users on each platform, which is more efficient? You might compare the promising user rate of different influencers to one another, since reliable metrics for influencer traffic can be hard to come by. If you’ve just started buying on Pinterest, you might use the promising user rate as an early indicator of how well this traffic is performing, before you have a steady flow of new and returning Pinterest users. If your promising user rate on a channel is particularly good, you might see how far you can expand your buy there until the promising user rate falls off.

No single metric tells the full story, of course, and by design the promising user rate is limited in scope: it tells you how the model assessed just the first session from new users. But this is nonetheless valuable data, particularly when combined with data from marketing platforms, or tools like Google Analytics. Prospecting channels are difficult to measure effectively, and the promising user metric is one more piece of the puzzle you can assemble to form a clear picture of your traffic.

When (and how) should I expect the report to be delivered to me? What is the date range?

Reports are sent by email on Monday mornings (EST) to whatever list of recipients you provide. You will receive one email with the report as a pdf, and another with the same data in a csv.

The first pane (“All promising users by source”) reports on the past complete week, where a week is Monday-Sunday. So if a Monday is on the 8th, the report will show data for the 1st through the 7th. The second pane compares the past complete week to the week prior.

What if I would like to make changes to the report?

Please ask! We’ve customized this report for a number of our clients to ensure that they’re getting the maximum amount of value out of their predictive model. One of our analysts would be happy to sit in on one of your check-ins and hear your feedback.

‍