I am forcing myself to write this bit. I am probably like you: I find analytics stuff incredibly boring.
It has also defined my career.
The thing is, if you don't build your startup around a deep understanding of analytics, you will likely waste huge amounts of time and/or money on the wrong things.
You need a direction to spend the limited time you have, and analytics is the guide. Without them, you are driving around without GPS in a place you know nothing about.
Let's begin:
what the hell is digital analytics anyway?
It's mainly gathering data on how users use your website and/or app so you can better understand what the user does. Specifically:
- Where they go
- What pages they linger on
- What pages they leave immediately
- what features they use
- How long they spend
- Where they came from <-- super important
- and a lot more
This means you can look at where users come from, such as your social media posts, and see if those users end up making a purchase or signing up for your product or service.
Analytics also helps an enormous amount with the A/B testing stuff we will be talking about later, so you can see if changes you make actually help users do things.
Story time: how travel buyers think
I noticed something unique about a customer of mine a while back: they had a lot of traffic to their website during the week, but almost no sales. The strange thing was that they had very little traffic during the weekend, and yet that was when they got the majority of their actual sales.
This turned into a conversion rate of something like 1.5% during the week and then 15% during the weekend (numbers are made up, but not far off from what the reality was).
I had to really scratch my head to figure this one out, as it simply didn't line up with anything I had encountered in the past.
More traffic can sometimes correlate with lower conversion rate simply because the traffic was less relevant (eg. big news came out about the company so people went to the site out of curiosity but wouldn't purchase) but then...
Wait.
_Relevance. _
Why was the traffic that came during the weekend more relevant than weekdays?
I took a look at the traffic sources and during the week, most of the traffic came from ads or social media. I also noticed that the users were almost entirely new users.
The weekend on the other hand was mostly direct traffic, which often means people typing the website directly into the address bar of their browser.
The weekend was also mostly returning users.
That's when it dawned on me:
During the week, people were at work and daydreaming about travelling. When they saw an ad or social post they would click it out of curiosity but wouldn't make such a big purchase while sitting at their desk at work.
During the weekend when at home and often after talking with their family, THAT is when they made the decision to actually purchase the trip.
We were able to work on a change in tactics (such as a focus on collecting emails during the week instead of focusing on sales and then using email automation to remind people to purchase on the weekends) which lead to a large increase in sales.
That's the power of analytics. Without solid digital analytics, none of that would have been possible.
Signal versus noise - not the blog
There used to be this idea that you needed to collect every possible piece of data you could.
The issue is that this generates a TON of noise.
If a metric changes, is that information useful? Can you action something based on it? Should you?
When there is a flood of data, it's impossible to know what you should actually do with it, or what actually matters.
The amount of data you have is no longer a strength, it's a liability. With data leaks destroying businesses and storage costs inflating, it is no longer viable to just collect everything and hope it'll be useful at some point.
Instead, you want to find a signal.
Data should be something that creates action: that is always the goal. If a metric changes, you need to know what levers you have to impact that metric and thus have control over what to do.
It should also be tied to a goal of some type. That goal matters a lot and should be the pillar you base all your decisions on. There should be only one goal.
Most often, the goal is money.
The hard part is that in digital analytics, it can be difficult to track money in some circumstances. Sure, if you are a retailer selling a product on your website, it is relatively simple.
What about services, though? A consultant like me doesn't charge via my website.
What about purchases that happen in person?
In those circumstances, I would recommend thinking about ONE step away from money. No more than that.
Often this comes in the form of a lead.
Wherever you land in terms of the goal, that needs to be your conversion/key event (word depends on which analytics tool you use).
The idea here is to avoid noise, especially vanity metrics.
Wait... What are vanity metrics?
Thanks for asking, Kole! I am happy to answer that!
A vanity metrics is a metric that feels like it is good but doesn't actually translate to your goal.
Let's say your goal is sales. You think, well, if I make some viral social media posts and get lots of likes, I will make more sales.
The issue is that likes from social media don't translate to sales. You can get hundreds of thousands of likes and have zero change in sales.
Here are some examples of vanity metrics:
Retweets/likes
Social media followers
Podcast listeners/viewers
page views
It might feel nice to increase these numbers, but they don't translate to sales as directly as you might think. You need to think in numbers closer to money.
Events, Sessions, Users and more!
I am not going to dwell too long on this bit as I promised I wouldn't get too technical, but there are some core concepts that if you understand and internalize early will help an enormous amount with finding insights in analytics.
Mainly, you need to know the difference between an event, a session and a user.
The problem that I encounter a ton is a simple question:
How much traffic did I get?
The issue is that basically no analytics tool has a "traffic" metric. They instead will have sessions (or visits if you live in Adobe world), page views (an event) and users.
Sessions
A session is a visit to your site by a device within a 30-minute time period (or so).
That is a very technical and exact definition, but that can help. What that means is if you visit a website from your laptop, go to 5 blog posts, then open your phone and go to the website as well, that will show up as 2 sessions. One on your phone, one on your laptop.
The number of pages you view does not impact this number.
If you stay on the site for longer than half an hour, the session will be split into 2.
When someone asks for "traffic" this is usually what they mean.
Page views
A page view is an event. This means any time a user opens a new page, it sends an event to Google or Adobe and is counted as one.
If you go to 5 blog posts in a session, like above, that means you will have 5 page views.
This is generally the highest number of the three.
Users
A user is similar to a session but doesn't have a time constraint. It's a way of counting how many individuals come to a website over a period of time.
If a user comes to a website, reads a bit, then leaves and then comes back the next day, that will be 1 users.
So using the example above of a person going to a website, browsing 5 pages, leaving, coming back the next day and reading another 2 blog posts. That creates:
1 user. 2 sessions 7 page views.
Here is a beautiful illustration:
One more technical note
I know, this is getting a bit much, but this one is important and something that has come up time and again.
There is a big difference between looking at successful sessions rather than successful conversions or events.
That's a bit of an awkward sentence but let me use an example:
Someone clicks a social media ad and then goes to your site.
They have to go so don't do anything on the site.
Later they see another ad and click it.
They click a few products, add one to their cart, and later make a purchase.
A few minutes later they see a deal and purchase that as well.
So as with the above, we look at the following:
2 ad clicks 1 user 1 session 2 add to carts 2 sales
Some places will use logic like this:
Count of sessions that click the ad divided by count of sessions that clicked the ad and sessions that include a sale.
If that's the case, it would be 1 session with both an ad click and a sale. You don't see the 2 sales that happened here.
This can massively change the reporting and show a very different conversion rate.
Let's say the numbers look like this:
100 ad clicks 85 sessions 50 sales 25 sessions with a sale
So if you use the session based reporting, you would get:
85 ad click sessions / 25 sessions with a sale, so a rate of 29%.
If you instead use events, you get:
100 ad clicks / 50 sales = 50% conversion rate.
Very different numbers. Which is correct? That's for you to consider.
The tools
There are many different analytics tools and the choice of which one to use is up to you, but I thought I would give a super quick overview of the main ones:
Google Analytics
The de-facto standard in digital analytics. This is mainly because it is free and has been around forever.
It is relatively simple to install, costs nothing, scales with your business and incredibly powerful. It also has a TON of support available online, mainly because so many people use it. If you need to customize it or help with something, you will for sure find support online.
I would honestly recommend you take a few hours to do the GA4 training you can find here. It is well put together and makes it really easy to understand almost everything I'll be discussing that is analytics related.
Adobe Analytics
You are a startup. Do NOT use this option.
I am including it here because it is relatively famous and has been around a long time.
Adobe Analytics is somewhat similar to Google's offering but has a lot less "out of the box". It is generally used by large corporations and is customized specifically for them.
Much of what you can do with it can be done with Google. You just need to know what an eVar or sProp is and how you can turn them into dimensions and events in GA4.
Kiss Metrics and all the rest:
There are lots of other digital analytics tools (like Kissmetrics) but I don't have much experience with them. In my years of doing this, I have only ever had one customer that used anything other than Google Analytics or Adobe Analytics.
If you want to explore them, go for it! But do be warned: the majority of the information out there is for Adobe or Google.
What about the built in data in Wordpress/Hubspot/Whatever?
It's... fine.
The data will be generally ok, but the caveats are usually that you have no way to customize it.
You can track your visitors, for example, but may not be able to track an ad campaign and see what sales came from it.
I would recommend switching to an analytics tool as early as possible in the life of your startup. That way you can have more data to work with as you go.
Okay, that's enough, let's move on ------>