Why the last mile of analytics matters
The last mile of analytics can take our state-of-the-art data stack and actually make it useful. Find out how.
Taylor Brownlow
08 November, 2021 · 5 min read

A typical consumer of data — one who uses operational dashboards, and regularly asks the data team for reports and figures — might be shocked to learn that our industry is emerging from a period of significant and sustained transformation.

Our data pipelines are more robust, and our data warehouses are now blisteringly fast making data more trustworthy and accessible than ever before. Yet, despite these innovations, the way our business partners access and apply insights remains unchanged. Even in the most data-driven companies where numbers are plentiful, working out what to do with those insights is sadly still a struggle.

As I see it, this is the challenge now facing our industry — improving the last mile of analytics to make our data not just more accessible but more actionable. To succeed will require a fundamental shift in our attitudes, technology, and leadership, but it’s truly the only way to share the potential of the modern data stack with the wider business.

From questions to decisions

Nearly a decade ago Gartner prophesized our focus on data-driven decision making in the analytical continuum.

While I disagree with most of the article’s assessments on how we are to achieve this, I wholeheartedly agree with the overall sentiment —the goal for every modern analytics team is to enable the business to make better decisions.

It’s easy to hear that and argue that we do support decision-making today. Don’t our dashboards help users prioritize their work? And don’t data apps help us evaluate trade-offs? But if we’re honest, we know the answer is, at best, sometimes.

When we get a request for data from a project team, we might think (and hope) that they will use this dashboard (or notebook, or chart) to challenge their assumptions, evaluate options, and spark creative ideas. That this data might be a beacon of light in the storm of a tough decision.

At Count, when we speak to consumers of data, they tell a different story.

Our data can be helpful early in the decision process when the questions are clear. How bad is our churn rate? At which parts of the funnel is it worst? For which users?

But as decision-making progresses, things change. Business users tell us that after receiving some initial numbers, they often either:

  1. filter questions to the data team because they are worried they will take too long, or aren’t sure they’re valuable enough, or
  2. don’t even know what questions to ask.

And that is the crux of the problem. Our business partners are not waiting for us to achieve Gartner’s pinnacle of analytical maturity before making decisions. They’re doing it right now with whatever info they have, and far too often, that does not include data.

Our waterfall development cycle cannot keep up with the rapid pace and bespoke nature of decision-making. Our tools do little to integrate data into the heart of the problems it describes. And perhaps most troubling, we’re so out of touch they aren’t even asking us for help.

A new analytical experience

In order to fundamentally shift this dynamic, we are in need of an equally significant change in how we work. Over the last year, I’ve been working with the wider team at Count to think through the last piece of the modern data stuck puzzle — the “last mile” issue. From our experiences building Count and our wider research we want to put forward the following three tenets:

1. Let the walls fall

Most data teams, intentionally or not, have created a wall between themselves and the business. Whether you’re an analyst embedded in a business function or part of a central data team, this is true. Requests are “lobbed” over to us, and we throw back charts, dashboards, and numbers.

I remember believing this wall was essential when I was working in a large company. The wall helped us maintain proper data security, prevented teams from coming up with their own warring versions of KPIs, and it kept us from being inundated with a mob of requests.

But now, our data is more secure and trusted, making the wall both unnecessary and downright problematic when we think about data-driven decision-making.

Dismantling the wall that divides us from the business opens up an entirely new way of working, one that is essential if we are to successfully integrate data into the business.

2. Beyond charts

In a recent post, Benn Stancil argues data, at its most effective, blends into the larger problem it describes. In those cases, “[w]e don’t immediately notice how much we use data…because data isn't detached from the rest of the experience.”

To get our data seen and understood, we will be called upon to present data in an entirely new way; one that is rooted in the context of the larger story, and one that is ruthlessly about the impact of our insights.

Current data tools fall embarrassingly short here. We present numbers, often very beautifully, but the burden of interpreting, applying, and acting on those numbers is up to our users. And that is simply too much to ask of them when they are in the throws of complex decision-making.

A new generation of B.I. tools are needed to fill this gap and give us a way to make the value of our insights more easily understood.

3. Embrace the chaos

Most of the work we do in data is in an effort to reduce entropy — we model data to remove inaccuracies, we turn commonly asked questions into self-serve reports, and we funnel ad-hoc questions into a formalized request process. This kind of attitude is in our nature as data practitioners and largely serves us well. But in the case of driving decisions with data, we need to challenge our instincts and embrace the chaos.

The most valuable way we interact with our data partners today is the messiest — it’s in the slack conversations, the over-the-shoulder demonstrations, or the ad-hoc email with a thrown-together chart with a few bullet points.

This work is both quick and messy, or what might be called agile in any other context.

Yet, we’re still operating in a waterfall development style which turns us away from these valuable interactions with our business partners. This way of working fails to capture the iterative nature of problem-solving with data, which most of us still see as an annoyance to eradicate rather than gold to be mined.

In the future, we will be working much more closely with our business partners in a way that encourages the rapid iteration intrinsic to decision-making.

Where do we go from here?

It’s clear we are early on this journey, but I’m heartened by recent concepts like the data mesh and data as a product that emphasize removing the barriers between the data and the business, and a more empathetic approach to sharing data respectively. It feels like our industry is slowly shifting its focus to the last mile problem.

Of course, there are still many unanswered questions — what does this future data organization look like? Do we even have one? How does the role of the analyst adapt to these new demands? etc. But isn’t that part of the fun?

I’ll be tackling some of these questions over the next few months in more posts. You can follow along here!

At Count, we’re building for the era of data-driven decision-making. Learn more about our collaborative data platform, or subscribe to hear more about the challenges and opportunities that lie ahead in More than Numbers.