White collar fail: total traffic and conversion rate KPIs
Why we should stop doing this, why we probably do this, and what we should do instead.
[acknowledgements before we begin — my web analytics foundation and this post were deeply influenced by the original web analytics guru Avinash Kaushik, in his Occam’s Razor blog post in 2010: here]
Since about 10 years ago, it’s had its place in individual goals, team targets, & financial reports: the total traffic & conversion rate% duo.
These metrics are secondary KPIs, at best.
Even executives fail to understand that online traffic is a spectrum, there’s always a level of quality attached— quality traffic: what we should care about and then there’s the rest: mistakenly-clicked, bouncing, will-not-come-back-to-the-site ever traffic.
How many times have we heard in a meeting:
“Well, traffic beat targets, but conversion rate% fell, so we were not able to hit the total target”.
Whenever I hear this, I think:
Why penalize your product team when marketing was forced to drive low quality traffic to meet goals?
Why not focus on the absolute amount of quality traffic and improving usability for users who could possibly become heavy users / good customers.
Rest of traffic is not worth much—usually the cost of acquiring these rest of traffic outweigh the revenue generated by ads — therefore not worth acquiring at all.
I’ve seen cases where a marketing team was throwing away millions in ad spend, and the result was avg bounce rate of over 80%(bounce: site visitors who leave after viewing only 1 page), just because the team had a total traffic KPI— this happens more often that you would think.
So what’s going on here?
My hypothesis is that, leaders are misinterpreting the online user behavior, and using the traditional retail: Traffic x conversion% x AOV calculation, simply applying it to online with limited understanding of the nature of online advertising.
The difference is that retail store traffic is basically window-shoppers, and these window-shoppers are somewhat interested vs online rest of traffic can be bought very easily, if the aim is to simply increase traffic.
We’ve all been there, right— not remembering what you clicked on to get to the page you have zero interest in— something that never happens offline, except when you’re drunk and in front of a fast food chain.
(but then again, maybe the neon sign is the siren that lures the drunken hungry tired core customers segment)
So the next best thing:
Instead of aggregates, why not define quality users (traffic) by deep-diving into who the core users are and determining which factors lead to acquiring these core users— for instance creating a segment of users based on page depth report, or those who add an item to shopping cart, etc.
Custom segments can be easily created in Google Analytics, and regular reports can be scheduled based on your own definition of quality traffic, instead of the total traffic so this can easily be tracked.
(For fast-moving goods ecommerce sites, the quality of traffic can also be defined by the average order volume of the purchase. The higher the AOV, higher the quality of traffic)
As for conversion rate%, instead of the aggregate rate% — which does not tell us much — why not focus on total # of conversion by user type: how many conversion is by new users, how much by repeats. (upcoming article on CAC and CPRO).
The absolute conversion# is more actionable — let’s say you realize that you’re trending to reach -10% of conversions for the month, then there can be a conscious decision of “do we spend more into marketing / push more sales to acquire these 10% conversions, or are we happy with the -10% and not to spend more budget and sales resources”.
Having said that, improving conversion rate% is important(Step 1 of growth hacking: here), but we must look at comparable channels & periods– such as conversion rate% of Direct traffic during similar off-seasons timing.
Conversion rate improvement is a longer process including service design that spans from user interviews, ideation, pilots and alterations over numerous prints, often times too much to “push” for that certain month.
Now, you can see how absurd it is to set an aggregate traffic target and a conversion rate% target?
— — — — — — — — — — — — — — — — — — — — — — — — — — —
If you found this article useful, share it with others:)
Or leave a comment, or send a message to my LinkedIn here
— — — — — — — — — — — — — — — — — — — — — — — — — — —
