tl;dr : The stages of my analytics evolution were, “Is anyone out there?”, “Stare and Hope for Inspiration”, “Event Obsession”, “AARRR Age”, “So What – the teenage years” and “Question Driven Analytics” and “Simple Numbers”.
What gets measured, gets done
– Peter Drucker
I read that quote years ago, and Peter Drucker is rich and has written lots of books, so I took it to heart. He is of course correct. Providing clear metrics on my performance, and the performance of others, is the easiest and most liberating approach I’ve found to working with a group of folk. Everybody owns a number. But, before I managed to make use of the data, there was a lot of wasted time. So here are the stages of my analytics evolution …
Is Anyone Out There?
Just like a newborn kid, I didn’t know much when I first installed the analytics tracking code into my first website. I stuffed the pretty graphs into my eyeballs like an infant satisfies their oral obsession by stuffing anything they can get their hands on, into their mouths.
It felt good. Staring at the number of hits was fascinating. Eventually I got bored with GA and moved on to Statcounter. The big difference here was being able to see the actually IP addresses of my visitors versus which GA anonymises the data.
At this stage, I didn’t even know that looking at the numbers was supposed to be a pre-cursor to actually taking action. Instead I used it as a kind of visual therapy. Reassuring myself that at least someone was landing on my website.
The Stare and Hope for Inspiration Stage
Eventually I realised that analytics wasn’t just there to make me feel less lonely in my lonely underground office where I saw no one from one end of the day to another (as it turned out, selling SMS gateway access in 2000 wasn’t a big money spinner). I’m so glad I got out of that business.
I got frustrated with the pretty graphs. They sat there laughing at my inability to turn their information into more money in my back pocket. I resolved that I just didn’t understand them enough. Like a crazed technical day trader I spent an hour or so each morning poking my way around the data.
No one told me that GA was really built for publishing businesses, and couldn’t really help me understand how to sell my SaaS app without some serious configuration tweaks. It was the wrong tool for the job.
It was a “stare and wait for inspiration” strategy that actually, at least taught me how to twiddle all the interesting nobs like filtering traffic by source.
If I were selling ad space in a publishing business, I need never really have gone any further. I could see which pages were popular, and I could see where the visitors came from to view those pages.
I did notice that download.com was sending me some traffic because I had uploaded a SMS Outlook plugin to their review site but for some reason this didn’t prompt me to go find other review sites to upload my software to. I had no way of knowing how much that link was worth to me. I had no goals or eCommerce settings configured. If I had, maybe I could have extrapolated forward and figured out how much review traffic I needed to become wildly wealthy or at least rich enough to pay the rent.
But I didn’t and inspiration never came.
The Event Obsession
Part of my struggle when it came to my analytics account was the raw volume of data. It’s too much of any mere mortal to parse and it sure as hell isn’t structured with a boot-strapped programmer with no marketing experience in mind. It was time to start eventifying.
Events are custom snippets of code that can be used to track specific occurrences. I think they were originally used to help track AJAX calls in an app that didn’t trigger page refreshes.
I used them for something else – to cut through the crap. By attaching events to specific clicks on the webapp, eg: the pricing page, followed by the sign up link I began to just focus on the events that meant something to my conversion funnel.
Events were also useful for pulling in data from other sources such as the phone system. I could begin to see which traffic resulted in time on the phone, which I knew resulted in sales.
I was taking back control – not just taking what was given to me.
I became more active in my thought process around what was important. Using events you can design the interaction flow you’d like to see, and then see if it works out that way.
I still mark out key events and track them rather than relying on URL’s.
The AARRR Metrics Age
I owe a lot to Pirate Metrics. Mostly because it taught me to stop worrying and love the data that mattered. The torrent of information is mostly useless and combined with my newfound love of events, it is fairly easy, even in GA, to start to organise the traffic into something more useful. Of course, GA’s media publishing background means it is still pretty lousy at helping you to track individual users.
For Pirates, track.io does the best job of helping to look at your users in terms of cohorts and (got forbid) real people you can interact with. Its built with AARRR in mind and you can assemble your events into the appropriate stages within the app without having to do too much thinking before hand.
The ‘so what’ Teenage Years
This is all great. I’m measuring my app like the fancy Silicon Valley investors recommend me too.
But hold on. How do you use this? After all this evolution, am I any closer to actually using my analytics setup for anything other than masturbatory aid to believing I was making progress.
Separating out steps in the funnel through Acquisition, Activation, Retention, Referral and Revenue was okay, but “so what”? If wasn’t AB testing each stage, calling up prospects / doing whatever it took to push people through that funnel, it was all for nothing.
The Question Driven Phase
So lets turn this on its head. Lets forget about code and talk about activities. The things that my analytics account measure are either changes to the app or the effect of marketing effort. Both take time. A thing I have a finite amount of.
So lets instead imagine a block of 100 marketing hours. What would really be useful for me to know about that effort. Lets take a simple one …
Which marketing hours created a customer?
Interesting. Now when I crack open my dashboard, I have a question to answer. Suddenly things aren’t so confusing. The colourful graphs can be ignored unless they help answer my question. This of course normally highlights woeful inadequacies of the setup to tack things like offline sales or even which of the marketing team were effective in generating those sales.
But these are problems that can be solved.
But that isn’t quiet enough. There has to be some impact on activities having found our answer. Let us say we find that most of our sales come from Quora comments. Do we have the ability to comment more or are we maxed out on the available topics we could comment on?
Which leads me to a more interesting question to answer.
Which activities don’t result in new customers?
Just as important as focusing resources on the things that get results, is freeing up resources on things which have no impact. Traffic sources (I’m looking at you Google+) that don’t result in customers but do take time to maintain need to be cut. Then it is a question of reallocating those resources to the activities that do get results. Or, looking for new activities that might yield results.
Enter, Lean Analytics.
Although the book claims to be about analytics, its really about metrics and how to come up with metrics you can use to take action. It emphasises simple metrics that everyone can grasp, that ties your team together. Those AARRR metrics aren’t a bad start. We started by sharing the numbers out to each member of the team. At weekly meetings we call out our number and discuss how we are going to nudge it in the right direction.
The AARRR metrics are a good start but they might not be the key metrics for every business. There is an art to figuring out which numbers really matter. For us, the number of leads, and the time on the phone are the key indicators. We also track the number of enquiries we are driving, and with those three numbers, we drive the business.
These are the numbers I pull out and slap onto a very, very pretty dashboard.
The fewer numbers on that dashboard the better. You can then do something clever like put them on the office wall on a nice big flat screen TV. Marketing can focus on their leads number, and sales can focus on the conversion. Everyone gets to see on a tidy screen together and everyone can see if a problem pops up. Eg: There aren’t enough lead or the conversion rate plummets.
Tools like dashing do a great job of pulling these numbers together.
Have I evolved?
So what is the next stage of enlightment for the analytics addict? Is there anything beyond question led analytics and communicable metrics or have I reached the highest level of enlightenment?