Not all conversions are created equal
February 2020
What should you be optimizing for?
Such an important question. Pick the wrong metric, and you end up expending considerable effort to end up worse than where you started.
But there are all kinds of businesses, all kinds of ways to make money, and all kinds of websites.
And for every type of website, there are metrics which, when optimized, bring little to no actual value. Let’s call these “💩 metrics.”
Then there are the metrics worth paying attention to - the “👍 metrics”.
And beyond that, in the realm of dreams and fantasy, are the metrics we’d love to optimize, but which present extreme challenges. Holy grail metrics, if you will - or just “🏆 metrics”.
Here’s a rundown of the 💩, the 👍, and the 🏆 for the 4 major types of websites - Ecommerce, SaaS, media, and lead generation.
Ecommerce
💩 Product Detail Page (PDP) views
Sure, you managed to get more people from the homepage to the product page. Yes, that is a necessary step toward purchase. But if none of those extra PDP viewers actually bought anything, you’ve just wasted their time.
👍 Revenue per visitor (RPV)
This metric captures actual economic value. And while getting more purchases could mean less revenue (because visitors are spending less) … or increasing average order value could mean less revenue (because visitors are making fewer purchases) … an increase in RPV is almost categorically a Good Thing.
🏆 Customer lifetime value (CLV)
Just as getting a visitor to a PDP doesn’t mean much if they don’t buy, getting them to buy is less than ideal if they never buy again.
The problem? Optimizing for this metric requires measuring how much each customer spends over a long period of time, and attributing their behavior to specific changes in the site experience.
(How can you know that a customer’s second boot purchase in 3 years had anything to do with the promo banner that was on the site when they first visited?)
SaaS
💩 Engagement
Similar to PDP views in ecomm, just getting people to browse around your site is meaningless if they’re not signing up for your product.
👍 Signups
Either this makes you money, or it represents a significant and necessary step in that direction. Creating an account and password is way more of a commitment than clicking a link.
If you offer a free trial, of course, your work is not yet done. But getting visitors to take this first big step is arguably the purpose of your marketing site - now your app and onboarding flow can take over.
🏆 Customer lifetime value
Say you’ve increased signups - and that even carries over to paid plan upgrades. Congrats!
But suppose these new customers end up churning much sooner than the rest of your users. You’ve got a problem. You’re a little richer than before, but not as rich as your experiment results seemed to indicate.
Successfully tracking CLV across experimental treatments will let you know what works best, long term, for your business. But just as with ecommerce, the timeline and reliability of tracking are problematic.
Media
💩 Time on site
While it seems like more eyeball-seconds is a Good Thing, this metric is typically unreliable (at least in Google Analytics). And as it’s an average, it can be skewed by a handful of outliers - visitors who spend tons of time on the site.
👍 Ad impressions, ad clicks, affiliate link clicks
Whichever of these actions actually makes you money … is the thing you should optimize for.
If your revenue comes from a combination of these inputs, estimating the value of each and producing a blended Revenue per Visitor metric is totally feasible, and will keep you from cannibalizing affiliate links in an effort to increase ad impressions.
🏆 Customer lifetime value
A pattern is emerging! Imagine you’ve made a change that causes an increase in ad clicks, without impacting affiliate link engagement. Awesome, you win!
Now imagine that these additional clicks are coming from visitors who are disappointed, even disgusted, by the post-click experience. So much so that they vow not to return to your site. Maybe they even post some snarky comments about it on social media.
Oops, you lose!
The ability to measure and optimize for customer loyalty would protect you from such a nightmare.
But, of course, it’s hard. People use different devices, in-app browsers. Some come back daily, but others might only return every few weeks. Unless you can install cameras in our houses, you’re probably stuck optimizing for revenue at the session level.
Lead generation
💩 Engagement
Just as with a SaaS business, getting people to click around your site may indicate an increase in interest … but it also might indicate an increase in confusion. Optimize for clicks and get … clicks. Who cares?
👍 Form submissions
This is the whole point of your site. If you can’t make a change that gets more visitors to convert to leads, why are you changing anything?
🏆 Sales qualified leads (SQLs), sales
We hope that more form submissions are a Good Thing, and that they’ll lead to more revenue. But plenty has to happen before that revenue materializes.
If these additional form submitters are unqualified, or spam, we haven’t really impacted the business. The true test is how many actual deals are closed.
The challenge? Deals take a while to close. They’re complex, and can involve multiple human beings - not just the one who submitted the form. Tracking all this back to an A/B test or marketing campaign can be impossible.
So what should you do?
Whichever type of site you have, there are 3 valuable steps you can take.
Throw out the 💩 metrics
If you’ve been optimizing for these, stop.
Double down on the 👍 metrics
Make sure you’re tracking them reliably, can segment them by device type and traffic source, and can attribute conversions to campaigns and A/B tests.
Dare to dream about the 🏆 metrics
Just because I say it’s hard to optimize for these metrics doesn’t mean you can’t figure it out.
A good first step is to ensure that you’re measuring your 🏆 conversion rate, and revisiting it every quarter.
Get to know this data - is it noisy? Does it vary greatly from month to month? Does there seem to be a single factor that influences it significantly? How might you test that hypothesis?
The transition from 💩 to 👍 puts you way ahead of the pack. Striving for 🏆 will keep you challenged and force you to get creative. Good luck, and happy optimizing!