Early in my web ana­lyt­ics career, I was pre­sent­ing to the exec­u­tive team at a For­tune 1000 com­pany. As we dis­cussed their con­ver­sion rate trends and ideas for improve­ment, one of the exec­u­tives said, “These rec­om­men­da­tions are inter­est­ing, but are they really nec­es­sary? Based on some bench­marks I recently saw, our con­ver­sion rates are actu­ally well above the indus­try aver­age. Why should we change anything?”

I could write a book in response to this ques­tion – but let’s focus on the bench­marks for now. The desire to com­pare your per­for­mance to oth­ers is sim­ple human nature – every­one wants to know if they are doing bet­ter than the other guy. And while this is a fairly com­mon and seem­ingly sim­ple request, it is not such an easy answer.

The real­ity is that bench­marks – in the world of web ana­lyt­ics — are fairly mean­ing­less. Why? Because com­pa­nies mea­sure per­for­mance in dif­fer­ent ways. Take some­thing as sim­ple as retail con­ver­sion rates. Some com­pa­nies mea­sure retail con­ver­sion as Orders/Visits. Oth­ers mea­sure it as Orders/Daily Vis­i­tors. And oth­ers mea­sure it as Orders/Monthly Unique Vis­i­tors. So when you’re look­ing at a bench­mark, like the 2.13% aver­age I recently saw reported from Shop​.org, what does that really mean?

Con­ver­sion rates also fluc­tu­ate from week­day to week­end, over hol­i­days, and in response to exter­nal fac­tors like cam­paigns and key­word buys. They also respond to inter­nal fac­tors like home page changes that can­not always be iden­ti­fied when look­ing at aver­ages. They also vary greatly by geog­ra­phy. So when com­par­ing your con­ver­sion rates to oth­ers, which time frame would you pick? Which pro­mo­tional period would you select? Would you focus on world­wide, or just US? Even if you have good answers to these ques­tions, you have almost zero vis­i­bil­ity into how your peers arrived at their con­ver­sion rate. Again, what does that 2.13% really reflect?

Due to these, and many other fac­tors, it is mis­lead­ing at best to com­pare con­ver­sion rates from one site to another. And this is why Omni­ture has refrained from pub­lish­ing indus­try bench­marks over the years – the chal­lenge is not with data col­lec­tion, it’s with data def­i­n­i­tions and stan­dards. I know the Web Ana­lyt­ics Asso­ci­a­tion has a stan­dards com­mit­tee so per­haps some­day the value of bench­marks will improve. But no mat­ter what hap­pens in the future, your best strat­egy will always be to focus on improv­ing your web efforts based on your own ana­lyt­ics data. That’s the secret sauce.