Rebuilding (Google) Analytics
Our unexpected story about growth
We started building Volument as a side project because we wanted to grow. Soon we realized that this particular problem wasn’t solved at all. In fact, it was downright broken.
First, we wanted to know if people like our front page. Do they stay or do they leave? And how bad is the situation – how soon they leave?
It all started with the realization that the bounce rate in Google Analytics was broken
Google Analytics defines bounce in a wrong way. For example, if someone consumes your content for half an hour, and then leaves weighting alternatives — it’s a bounce. So people who absolutely love your product are considered bounces. [1]
How can our analytics go wrong this bad?
Typically reported bounces are that 88% of search traffic and 71% for social media traffic [2]. Are these are big claims real? How are they measured?
And if bounce data is broken, can you really trust the data that depends on it. Doesn’t all growth depend on it? People are either interested in your product or not, right?
Why aren’t we talking about this? How is it possible that this fundamental metric isn’t studied to death?
Google Analytics bounce rate is not just an analytic. It’s an agenda amplifier. Bring your agenda to the room and we’ll support it with data we got.Founder of User Interface Engineering
Perhaps our analytics is better at understanding people who stay.
Oh boy — how wrong we were.
Engagement
As writers, we are tremendously eager on knowing how people read our content.
How many read the article fully? How much time did people spend on it? Which sections are read the most? How about the crucial first-time visitors on our front page? A difference on mobile devices?
Again, no tool could answer these essential questions. Nobody could tell us whether our stuff is interesting or not. No analytics understood the time people spend reading, typing or scrolling.
How is this possible in 2018?
It is — and it gets worse.
Retention
Our biggest surprise was retention – the “king of all metrics”. At least that must be examined to death. But you guessed it. It’s not.
Do they come back? How much? Who returns the most? No luck.
We used Mixpanel to collect events. They trained us to ignore page views and focus on more important events, that are specific to our business. So we started tracking signups and purchases.
Then we saw that our retention data in Mixpanel is useless
But of course, this means that we only knew the retention numbers from people who registered or made a purchase. But honestly: how many times you register on a site when you browse the Internet?
So we knew the retention figures from maybe 2% of our audience because most just read and go away.
Mixpanel specifically advises us to ignore page views because according to them they are “the bullshit metrics”. And they still recommend this approach.
For them, retention is clearly not the king of all metrics.
(And our minds were not trained to understand these “cohort reports”)
Conversions
We really didn’t want to be another site that collects clicks and shoot popups on people’s faces.
We want to treat our visitors well and respect them as a person. We try to offer something that people actually want and not force them to do anything. We try to increase their willingess to convert so that the whole process happens naturally.
But of course, the current analytics tools don’t understand this kind of pursuits. They don’t care about human emotions. They simply focus on the outcomes: the signups and purchases, but ignore the cause and whether people want to convert in the first place.
Then we faced another meaty surprise — our conversion rates are also broken.
This cannot be real? We thought that the math goes normally, so that you take a bunch of people, wait for some time and see how many of them converted. That’s a rate, right?
Not at all. At least not in Google Analytics or in any other analytics tool we studied. The visitors are divided with another entity, like page views or sessions.
This was a big problem for our growth attempts. For example, our rates went down when we acquired new people on our site. Basically the better we are at marketing the poorer our rates are.
How rewarding.
The most important thing
We spent over a year just figuring out how everything in growth is tied together before doing any programming.
While studying, thinking and re-writing our guide we experienced two special moments. The first one was realizing the importance of time and how it is the only way to measure bounce rates or how much people engage with the content.
After we found our answers to user experience, engagement and retention we suddenly had the most desired metric on our hands: the product/market fit. And it turned out to be a very simple thing: you just need to study how visitors behave on their first visit and on the days that follow.
The only thing that matters is getting to product/market fit
Growth analytics
We build Volument for ourselves, so we could use our time on growth and not waste it on the wrong things.
Now Volument is absolutely invaluable for us. It tells us where we are, what are our bottlenecks, and what should we do next for achieving the maximum growth.
In fact, Volument helps us so much that we stopped worrying about money.
Because now we a more important thing to worry about: the product/market fit. We know that increasing this number all the rest, such as the revenue — will come naturally.
And now we know exactly how to do that.
Welcome to growth. Let’s fix it together.
Leave a Reply