June 26, 2019
“Partial Visitors”: or How Everyone’s an Expert
Recently, I heard of a situation in which programmer referred to visitors to the site as “partial visitors.” This was in a meeting with another marketing agency that works in analytics, and the business owners.
I was immediately livid at the connotation, as well as the assumption of analytics knowledge.
Maybe I should explain further that the programmer (we’ll call them Peanut) started because they didn’t want GA tracking code placed in the recommended spot (the <head>). He wanted it to load much later, after all of the site had loaded. Peanut then explained that because of the images, video and gifs they use, the site is already slower than the benchmark.
After admitting that they can’t speed up the site because of the content every other site has, they decided that the GA tag was where they would cut speed down.
When the issue of how this would affect the analytics, Peanut explained these “partial visitors” are those who click on the site and then leave because they made a mistake (like a bounce?). He then goes on to say that these are NOT useful to our analytics metrics and that we only want to track users that have landed on the site intentionally.
If you aren’t as livid I as I was then maybe I need to delve deeper into this and explain my reasoning (which you may feel free to disagree with)
[TL:DR version – I want to fix why people are bouncing, not ignore that they are.]
Issue #1 – Not Following Stated Recommendations
Whenever you buy new software, or a new piece of equipment, what does it say about modifying it? It voids the warranty. Why is that? Because you are using it in a way it’s not designed for.
In the same way, GA is saying that this is where the code needs to go for Tracking purposes. Of course they know speed is important, they rank sites based upon it. Why would they create a tag that would affect their own rankings? Obviously GA’s code is not perfect (far from it), but there is extensive testing by both GA engineers and the myriads of analysts around the world.
When you, as a developer, say you know more than all of those people who are constantly trying to break GA on a daily basis by pushing it to it’s limits, you are committing an arrogance of amazing proportions. You shouldn’t even be used as a developer because you can’t even realize that you aren’t an expert.
There are legitimate reasons why it may be necessary to push the code out of the <head>. There are companies that have to do it sometimes. But hopefully they were made aware of the consequences and were satisfied with the trade-off. But most likely they were told by IT that there was no issue, and that they don’t need those visitors anyway.
However, most of the time, we should just place the GA tracking code where it is designed to go.
Issue #2 – Not Listening to Marketers / Analysts as to Why Missing Data is an Issue, and Instead Claiming it Isn’t
If you were in a room with Stephen Hawking, and he started talking about physics, would you stop and correct him?
No, you wouldn’t. Because he’s an expert.
Why then is a developer (who may be an expert at what they do) telling the analytics expert they are wrong?
It’s disrespectful, rude and just plain unprofessional.
Issue #3 – Devaluation of Visitors
Data is the new oil. It is growing in value, and we are always looking for more. Always more. So then why would I not want to have the data on those people who came to my site then left immediately? This is normally called a ‘bounce’, which is very important to measure. Analysts do it all the time.
These visitors actually made it to my site, which makes that data worth more than most any other external source. They were so close to buying, reading, posting, or whatever it is I want my visitors to do. Closer than anyone else on Earth at that time.
And Peanut says they aren’t important?!
Devaluing the visitors is basically saying that this group is not as important as another group of visitors.
Peanut called them “partial visitors” because there was no understanding of them as people. To Peanut, they weren’t worthy of tracking, as if their actions are any less value to a true analyst.
They are “visitors” pure and simple. Nothing “partial” about them. They may not actually need to be on your site, and many of them probably were mistakes, however, that does not change the value of the data, nor the information that can be gleaned from their “mistake.”
Issue #4 – Misleading Reason Why They Shouldn’t be Tracked
Issue #4.a – Misleading Reason
“We only want to track users that have landed on the site intentionally.”
Well yes, and no.
If we are looking at a process funnel, whether an insurance quote, a t-shirt sale, or anything that involves steps for something like a transaction, then no, we don’t need those “partial visitors” IN THIS ANALYSIS!.
Unless… (big unless), you are measuring a landing page funnel. Where you will have a lot of bounces. This time you want them because you want to understand WHY they are bouncing so you can fix them from even coming in the first place.
If you send out an email campaign, people will click through, then bounce on your site. Why? Look in your analytics. Fix it. Lower that bounce rate. But you can’t do that if you aren’t tracking it in the first place. Don’t ignore it, just figure it out.
Issue #4.b – Reasons to Track
If we only looked at those who came to the site intentionally, then we would a 0% bounce rate. Because they would all have been on the right site the first time. No need to leave immediately. In which case we can go home. Our job is done.
However, if you happen to be one of the 100% of businesses that have a bounce rate above 0%, then you do want to know about it.
What if these “partial visitors” are coming from a Google Ads campaign? Something is wrong with that campaign, and we are losing money on paying for clicks that aren’t relevant. If we aren’t even tracking these visitors in the analytics, then the ads may be costing massive amounts of money with nobody noticing.
The same for an email campaign as mentioned above. You are paying to send those emails to many people who are already your customers. Don’t you want to track if they are bouncing, so you can fix what needs fixed?
Peanut’s reason for not tracking, is actually why we should track it. Unintentional visits are just as important, and sometimes even more financially important than intentional visits.
Issue #5 – Impersonating an Analyst: Or Someone With a Hint of a Trace of Analytics Knowledge
Seriously. Stick to developing, and I’ll stick to analysis. Yes, we need to break down silos and have more cross discipline knowledge, but remember that you aren’t an expert in every field, just your own. Allow other people to be experts in their field.
Ultimately situations like this come down to selfishness. In my view the Peanut is selfish because they don’t want to have to speed the site up, but would rather blame GA. Analysts can be selfish because they want as much data as possible, even if it means it could slow the site down.
The business owner is selfish because they want to run a business and make money, and are also the only ones allowed to be selfish. The developer and analyst must decide what works best for the business owner, not themselves.
In this case, more data is crucial to finding why people leave so suddenly. The speed of the GA tag isn’t bad enough to move it.
Follow the recommended guidelines. If you can’t, then have a good reason not to. Don’t use some crap excuse about something of which you know nothing.