Andrew McBurney's Blog | AgeTech and AdTech in Canada

In smaller companies, wearing many hats is the unsurprising norm.

In a technical capacity it's not uncommon for software engineers to handle database administration, application hosting and deployment pipelines, application architecture, security, analytics, quality control all on top of their actual main priority of writing code.

But this list is too short, in my opinion.

They should also be the go-to team members for measuring user behaviour.

The reality is... in a small team, the developers have insight and access to data that the company would otherwise never be aware of.

It all comes down to one small aspect of the broader data analytics and user experience disciplines: user behaviour data.

With basic segmentation, developers can uncover opportunities that drive product strategy, improve UX, and increase revenue.

The Year it Clicked

In 2018, I was busy running a service based business working with mostly other service based businesses, and helping them with their web needs. Building websites, setting up tracking, optimizing their website code to get more exposure in the search engines, connecting their website to their existing software and the like.

But I had one constantly nagging problem: all my best clients were exclusively using qualitative data to drive their sales efforts. They didn't actually know what was working, or why.

I knew this problem needed solved, and I dove in to learn how to fix it.

After researching, I signed up for ConversionXL's "Data Analytics Mini Degree" online course.

And my entire world view changed.

For the first time, somebody broke down how to tie raw metrics back to real world useful outcomes that are specific to business objectives.

Instantly, my mind flooded with ideas of how these concepts could be used, and it immediately allowed us to identify areas of opportunity to boost the number of customers the businesses would get without adding more advertising spend or team members into the mix.

I was hooked, and my love for a marketing concept called "conversion rate optimization" was born - a discipline I still study to this day.

The lesson was clear: all actions on websites generate metrics, and that data tells a story of "user segments" or cohorts. And those stories can be used to maximize outcomes by altering internal company efforts or focus.

Segments as Engineering Signals

Back to the technical end of the spectrum, let's "connect the dots" and see how this is useful for software engineers in small orgs.

Because the reality? Everything I explained in the previous section is more of a "marketing" function, not technical, right?

Kind of.

The reality: when a company is in the "figuring it out" phase, and processes aren't established yet, technical people become bridges to identify potential metrics or behaviours. They know if and how certain things can be tracked, and have the ability to ensure things are implemented correctly and strategically.

When developers know how to identify these areas, and understand its their job to be on the look out for opportunities, then creating pools of useful data becomes ingrained in the company.

These data pools then can be sliced and diced to become structured data segments as needed, and used in product decisions.

Identifying Encoded Information in User Segments

One of my earliest wins involving identifying a user segment and challenging a status quo assumption in a company was in 2019.

I had been brought in to help with the technical side of marketing operations. Essentially, un-silo data, and connect systems so we can make better choices. I dug into absolutely everything, and one interesting thing surfaced...

Their digital "point of sale" experience had a low success rate for people on mobile devices.

I asked questions to around this to relevant internal stakeholders, and the general company narrative was "Our customers don't buy on mobile, so we don't optimize for it."

Now, keep in mind... this company was very good at what they did. Their marketing talent far exceeded my own, and virtually every aspect of the company was working as a well oiled machine. It was profitable, lean, had loyal teams, happy customers and a solidified revenue model. This wasn't some "figuring it out" company. This was a real business.

Why did they think that users don't buy on mobile?

Because they had one raw metric that gets dumped to their sales platform that listed the device used for successful checkouts. Almost all of them were laptops of computers, with little to no smart phones or tablets being attributed to sales.

I found this odd because it did not align with my prior experience of mobile users buying on their devices, but I wasn't sure because of how sophisticated the company was.

So I got to work, I set up the basic tracking needed to identify the cohort that I wanted to measure: specifically, users on mobile devices going through the "point of sale" experience. Tracking what that segment of users were doing when landing on a sales page, clicking a buy now button, interacting with the checkout form and what not.

When I started putting the pieces together, it became very clear mobile users were trying to buy... but were getting frustrated or confused and abandoning the process.

Some of them came back on their computer or laptop and checked out successfully - but other sales were lost forever.

This is where the idea of seeing the "encoded" information comes into play.

Had I assumed that little successful mobile checkout metrics meant no user interest, I wouldn't have uncovered this treasure trove of new revenue. There was information baked into their internal reports that I was able to use to identify the "why" behind the missing mobile sales, versus accepting the raw metric as the "be all end all" of the situation.

In the end, it turned out to be a trusted and skilled internal web developer driving this narrative in the company.

The developer constantly echoed "Nobody buys on mobile, its a waste of time to focus on it" which meant no checkout flows or conversion paths were optimized.

One of the more obvious examples of this was how simple mobile friendliness was handled on the checkout page. On a computer, you visit the checkout page and it's immediately obvious you enter your details to checkout. This signals to the user they are in the right location, and can successfully check out. On mobile though? They had to scroll down about 5 or 6 screens to find the checkout. The text on the sidebar on desktop was showing first, making it not obvious it was even a checkout page. The title "Checkout" wasn't even displaying to the user upon loading the page.

We were able to run a split test using VWO and saw an immediate 19% boost in successful conversion rate for users going through the checkout flow because now mobile users could actually checkout. That's not chump change in a company that size.

Turning Segments into Action

I do not like hoarding data for the sake of it.

I do like collecting data that can be used for an obvious use case.

This is where identifying segments come into play.

Being able to identify potential segments of users that relate to certain ideal outcomes, and then analyzing and creating experiences for them to benefit is an easy win, assuming they're tied to business outcomes.

In a future post I'll dig deeper into this concept, and drill into real world examples through the lens of refactoring existing software interfaces to better connect to business outcomes, by analyzing real world data.

A Final Thought

Identifying user segments is not the job of a dev. It's the "business end" of the company.

But understanding these segments, and how they are flowing through the application?

That's where the dev shines.

As all things software - it's a team sport.

Dial in communication and ensure everybody is on the same page if you want to begin using the information hiding in plain sight to boost business outcomes.

User segments are not just analytics artifacts. They’re encoded behaviour signals. Developers who understand and act on this data significantly improve product outcomes and company growth.

Key Takeaways