The Four Stages of Data Maturity in SaaS

I recently spoke to a founder about how to approach his company’s data strategy. There are four key stages to a company’s level of data maturity.

My conversation came about because the founder wanted to know what my experience of evolving a company towards a more data-driven mindset consisted of. Changing the extent of data available to a company is hard, but changing behavior towards using data as a non-negotiable pre-requisite to any decision-making is harder still. But at least recognising the stages of data maturity is a good start.

Stage 1 — Gut Feel
When you first get started, you don’t really have much in the way of data. And so part of the founder’s magic sauce is their ability to check in with their gut and see what it is telling them.

“Will my customers prefer this feature or that feature? Let me just check what my gut-feeling is on this for an answer”.

Because gut feeling is based on all of your experience to date, it’s not totally crazy to consider it in the absence of data. Where gut feeling is super useful, is as part of the creative process. You need to be able to take what information you have and divine something new from it. And if you think about it, the early stages of a startup are all about creating something new, possibly with an angle that hasn’t been considered before.

The obvious problem with it is that just because you think something, or even the people you generally hang out with think something, that doesn’t necessarily translate to anything statistically valid. As they say, “the plural of anecdote is not data”.

Stage 2 — Data, data everywhere, but not a drop to drink
To (mis)quote from The Rime of the Ancient Mariner, at the second stage of maturity you have a load of data, but the ability to do anything with it is limited. Why? Because it is generally a very manual process to get at it and it’s siloed.

The likelihood is that you will have different tools for tracking different types of data. Examples include:
- Product usage
- Customer behavior (for example, expansion, contraction, renewal, churn, NPS)
- Revenue (growth)
- Sales funnels
- Marketing funnels
- Traffic data

Possibly you can look at some of this data by customer type (for example size of account, size of business, location etc), but it takes time to go and get the data and manipulate it yourself in a spreadsheet.

This is actually not a terrible thing because as a founder you will get a really good feel for your business’ key metrics, but it won’t be easy to do it at any scale. It will be time consuming and very probably a lot of it will reside in your head, however diligent you think you are at sharing. Even though creating reports is time consuming and manual, at least you are now able to make data driven decisions. Validating your ideas using data is infinitely better than leaving it to gastrointestinal superstition.

Understanding what is really happening with your customers rather than guessing, is going to allow you to improve the way you market, charge and interact with them, but it still feels like you’re not able to see the bigger picture.

The silo aspect of the data is problematic. You may be able to see how well a particular piece of content was in driving traffic to your website, but how easy is it to track the impact right the way through to becoming a customer and beyond?

Stage 3 — Central Data Platform
As you grow, you’ll find more people in the company desperate to get hold of data to measure what they are doing. Each department’s ability to extract and manipulate the data may vary and may just come down to the skills and willingness of particular team-mates to roll their sleeves up and get stuck in. This is where having a central data platform starts to become important.

Business Information tools pull in data to one unified platform and allow companies to get a broader sense of how the data connects up. It does this by connecting with your primary data sources (see the bullet point list above) and making it ‘easy’ to visualize that data. Suddenly you can create a dynamic report that overlays data from your CRM with data from your traffic analytics with data from feature usage on your product.

Inevitably, to do this you have to pay for your BI tool in the first place: tools like Tableau, Microsoft Power BI and Looker. These aren’t cheap. On top of that, unless you have those skills in-house, you’re likely going to need to find a Data Engineer to get things set up. As a result, Stage 3 isn’t normally reached until you get to a point where you have enough data and enough different people screaming for answers to data questions, and you can afford the investment in both the software and the expertise to make the data actually work reliably.

Inevitably what happens is the Data Engineer creates the most urgent reports demanded by the business and uses that process to bug test the integrations/connections to the primary data sources. This is an improvement on what came before, because the data is now connected up, but it still requires one further step top make the most of it.

The main thing here is that by the end of the process, the business should be reassured enough that the Data Platform can be relied upon as the Source of Truth.

Stage 4 — Data as Insight
Once the data is all firing and seems to be accurate, the final stage is when you move to having a Data Analyst on your team whose job it is, firstly to create and manage the various reporting that the business needs. Ideally this is more about creating reports that can be interrogated by individuals, rather than being on the receiving end of a load of report requests that they have to fulfil. But secondly, and perhaps this is Data Nirvana, actually use their skills to be able to spot insights in the data.

Along with an agreed set of reports that can be interrogated by anyone in the company, what you really want to end up with is someone who says something like, “I’ve been looking at this set of customers and I noticed that if they do x in the first few days after trial, they are ten times more likely to convert”. Identifying trends like this can give the product marketing team a better sense of what behavior to encourage new trialists towards for success. But it can also inspire the product team to reimagine some of their onboarding.

Even if these insights only create micro-improvements, if you can action several of them every month, the impact can be significant over time. Much more reliable than your stomach, however well-meaning it is