S2 E1: Solving Self-Serve Analytics

In the last and final episode of the Building CAUSL series, I shared the news that CAUSL was merging with LeanConvert and I was joining as their Director of Experimentation and Analytics. I finished that newsletter with a Marvel-style post-credit teaser stating that whilst I'm closing down the consulting side of CAUSL, I am building CAUSL Analytics - a free self serve analytics toolkit for analysts and non-analysts. Screenshot of the currently available tools at the end of this post.

Why Solve Self-Serve Analytics?

Admittedly I should probably spend my time solving problems that actually matter, such as world hunger, but that's an insurmountable problem and I don't even think I'd make a dent in solving it. So I'm going to spend my time solving the problem that has haunted me my entire career - self-serve.

Definition Of Self-Serve Analytics

Before I answer the question of why I'm trying to solve self-serve analytics I should start with a definition of self-serve analytics. Self-service analytics is a program of work that allows non-analytical users inside an organisation to access and analyse data without relying on an analytics team for support. In some cases this is as simple as granting access to data and tools, but in most cases it requires some training. Platforms such as Google Analytics and Amplitude already do this to a certain extent with a certain type of data set (I won't go into this now). But as great as they are, there are certain things that they don't allow you to do.

Biggest Failure Of My Career

I've been trying to solve the Self-Serve Analytics problem for more than a decade now and I've failed every time. I should state that I have never failed to launch the program or even get people to sign up and attend some training sessions. The problem I've faced and where I feel like I've failed is that like most non-essential programs inside an organisation, it fizzled out after a while. Why? I don't know. I suspect it's because the programs lost momentum due to more pressing issues that needed to be solved; and let's face it, there are always more pressing issues. In a nutshell, despite my best intentions, I never gave the program the attention it deserved and I never allowed it to be a P1 priority.

Why I Think I Can Solve It Now

I don't, but I think I stand a fighting chance now that my mission to solve this is no longer aligned with an organisation and I can give it the attention it deserves. There are no KRs against it (maybe one), no quarterly deadline to achieve this by and there are no other competing priorities - just my work, my family, crap talks, and my social life - although that last one is less of a problem these days.

Why It's Important

Now we get to the why! My raison d'etre. If product analytics and experimentation are my passion, then solving self-serve analytics is my purpose because I believe that anyone can be an analyst.

I realised a long time ago that 80% of the work my team was doing could have been self-served by the stakeholder. I believe a successful self-serve program would have empowered users to get the insights they needed to make informed decisions quickly and easily without the bottle neck of an analytics team. This would have in-turn left my team to work on problems that would have had a much bigger global impact to the organisation.

How I plan to solve Self-Serve Analytics?

Who Am I Solving This For?

Analytics is a broad subject spanning many, many different areas. So to help you (and mostly myself) understand how I plan to solve this, I start with who I'm solving this for? My audience is the non-analyst - the product manager, the marketing manager, the conversion optimisation specialist, the engineer, trading manager, the startup founder, the dinner ladies - ok not the dinner ladies but you get the idea. Anyone who relies on an analyst for support. I don't need to solve this for analysts and data scientists - if anything I'll probably enlist some of them to help me on my mission. Nor do the solutions need to be around complex analytical methods. The every day person does not need to know how to train a machine learning model, but they might need to understand how many people to include in their experiment.

The Motivation Hurdle

One of the things I should have mentioned in the "Biggest Failure Of My Career" section of this post is that I might have failed because there is an intrinsic lack of motivation to self-serve. Maybe people just don't want to "do analytics" themselves. I don't think that's the case though. Below is a short survey I did on Linked asking non-analysts if they would do "basic self-serve analysis". I know this is not how you conduct proper research but given that most of my non-analytical followers and connections are my target audience, I think it might be a good enough indicator of my overall target population, at least at a surface level. I need to dig into "why they don't currently do it themselves and what's stopping them?". If you are the target audience, please could you answer these two questions in the comment section of this post?

How I Plan On Solving Self-Serve Analytics

Which finally brings me onto the "How?". I've already mentioned that I've built a number of tools as part of CAUSL Analytics to address this - some of these tools exist in isolation out in the wild and others are behind paywalls. I'm just consolidating everything and more. But access to tools without knowing how to use them is not very helpful. So the next phase of this will be education and training. I'll be creating training videos, posts and seminars around the tools, the different input variables, interpretation of results, challenging assumptions and outputting the results. I even plan on having a 10 minute tutorial section during the CRAP Talks events ( Parveen Downer / Becky Lacock - FYI) and leveraging experts in the industry to help me with this mission ( Shaun McGirr , Tris J Burns ?? , Nicolae Chelea , Lottie Linter and ???? Julie Hoang to name a few that I have yet to convince). If you want to join me on this mission, please do reach out.

Wrap It Up Now

Measuring Success

I mentioned that I won't have OKR's or deadlines, but I figured I should have a north star metric in mind and that is going to be Monthly Active Users - how original! I figured I'd try and get to 3,000 monthly active users by summer next year which is ambitious but hopefully not outrageous. The website has been tagged now using PostHog which is free and awesome so I can report on the numbers transparently. You can see on the first graph below that I have a long way to go.

Right, I've rambled on for long enough now. I'll finish by saying that the platform is live, so please feel free to sign up. It's free to use. If you'd like to get involved in my mission to make self-serve analytics a success, please reach out.

Thanks for getting to the end. More to come on this in future episodes.

Bhav

(Screenshots below)

PostHog Dashboard Screenshot

Screenshot of CAUSL Analytics