More thoughts on using visits or visitors to calculate conversion rates
Recently I was talking to a friend who was asking about my post on buyer versus order conversion rates I posted recently. We had been talking about the “every session is an opportunity to convert” mantra that some folks push as gospel; his comment to me was funny. He basically said, “I manage analytics for a company that does over $100 million annually through our online channel and that type of thinking is [crap].”
I told him to tell me how he really felt.
After he read my post he said he’d started calculating the delta between buyer and order conversion rates for his own site on both a daily and monthly basis; he’d been calculating both buyer and order conversion rates as part of his daily KPI set but hadn’t really thought about the difference between them. While he wasn’t surprised to see an average of six to eight percent difference on a daily basis, he was surprised to see that on a monthly basis his order (visit-based) conversion rate was, on average, twenty-seven percent lower than his buyer (visitor-based) conversion rate!
Put another way, by subscribing to the “only use visits to calculate conversion” methodology my friend would be under-reporting the likelihood that he would sell products to real people on a monthly basis by nearly one-third!
So he got me thinking, I wonder what the monthly delta is between buyer and order conversion rates (BOCR Delta) is for book sales on my web site. Have a look:
Aside from the fact that conversion is off slightly over the past few months, likely owing to the fact that I’ve stepped up my efforts to bring traffic to the web site, you can see that I have much the same problem as my friend on a monthly basis. Were I to rely on visit-based conversion rates alone, my understanding of how real people purchase on my web site would be incomplete.
Anyway, I stand by my original statement, you need both visit- and visitor-based conversion rates to understand how your audience converts. Both metrics tell you something valuable; one tells you about the person doing the converting, the other tells you about the process.
I welcome your comments on this subject. Perhaps you disagree with me? Or perhaps you agree but are having a hard time calculating one or the other rates using your web analytics application?
Anonymous added the following ...
Hi Eric,
I think this speaks directly to the concept that it normally takes more than one visit for people to convert. This was a point brought up at SES San Jose in the first metrics session–that it takes something like 3.6 visits to convert (based on their study which should not be assumed to be universal truth). But thanks for putting this up. It’s a great reminder that visitor behavior is complicated and takes special care to understand.
Jason
Read my blog here
Eric added the following ...
Anonymous: Perhaps, but this would be a good place to add segmentation to your conversion rates such that you could examine the OCR and BCR in the context of “existing customers” and “new customers” … I would expect, in the situation you describe, the OCR and BCR would be very similar for existing customers (they already know the site and the purchase process, right?) and likely divergent for new customers (depending on the level of consideration associated with the purchase or transaction.)
For content and information sites the BCR will describe the propensity to get information into people’s hands while the OCR will describe how easy it is for them to get that information.
Excellent points!
Eric added the following ...
Jason: Not a problem! I mean, on some level I do get why people say “every visit is an opportunity to convert” but in my experience when money or personal information are involved, visitors are more cautious and it may take several visits for the conversion to occur.
Who’s study are you citing? I missed the panel …
Thanks for your comment!
Anonymous added the following ...
Eric: I believe it was the 360i presentation on the first day in the Search Behavior Research Update session. (someone correct me if I’ve given credit to the wrong presenter). They had broken it out by natural listings, paid listing, overall and combined. Combined it was 3.9 clicks to purchase (conversion). This was using a traditional e-commerce model. Other less rigorous types of conversions may have higher click to purchase ratios. However, the visit to visitor ratio is relevant and important for all types of conversions.
Anonymous added the following ...
Agree on the need to look at visitor and visit conversions - however, dose’t it also depand upon the type of business? For example, a purely transactional business where the same customers generally make repeated purchases (a majority of the sales) I would think would rely more on the visit conversion while a content/information site would be more likely to use a visitor conversion.
just a thought