“Big data” is all the buzz. It’s a challenge to merchants and data providers alike. Complex calculations on large volumes of data can be fascinating to a programmer. But these aren’t just numbers – they represent real-life actions of people on a website. If we aren’t careful, our calculations can outsmart the spirit of these actions, resulting in meaningless (or at least misguided) data for merchandisers.
Recently we undertook a project to create a new dashboard for the SLI Merchandising and Reporting Console. With more than 100 metrics already available in our console, it would have been easy to present some of these and call it a day.
But we decided it was time for a fresh start. We talked to our customers about what was most important to them. Then we went further – all the way back to the basics. We examined our underlying calculations to ensure the way we tracked data considered customers on their shopping journey and not just purchases.
Many existing reporting systems, including our own, count events like transactions and compare them to other events like page or session counts. But tracking events in isolation doesn’t always reflect how shoppers behave. Consider a shopper who starts browsing in the early evening, breaks for dinner and returns to make a purchase after putting the kids to bed or the shopper who has a late-night browsing habit and completes her purchase after midnight. Many analytics simplify and examine activity based on a set time period. As a result, each of the previous examples would likely be considered as two sessions – one that failed to convert and one that succeeded – when in actuality each was a single shopper.
A session-focused versus shopper-focused model can cause e-commerce sites to worry needlessly about conversion rates when in actuality the trouble is with how their analytics views the data.
Another example of analytics skewing reality is when a customer completes a transaction and then realizes he forgot something and makes a second purchase. Should one customer be treated as two just because he completed a second transaction? When considering conversions, is a shopper who transacts twice buying one item each time really more valuable than the shopper buying two items in a single transaction? We don’t think so.
The new SLI Dashboard now includes shopper-focused attribution. As a result, SLI looks at all of a shopper’s activity in the 24 hours leading up to a transaction, regardless of when it occurred. A site visitor with activity before midnight that continues into the next day is counted once, and multiple conversions from a single customer are treated as one successful conversion.
Conversion rate and average order value are critical measures of your e-commerce site. You need them to be an accurate representation of how many people successfully found what they were looking for, how many items people bought and how much they spent. The new SLI Dashboard provides a real-life view of how your customers are interacting with your site. In a world of numbers and calculations, let’s remember our customers are people too.
