What if analyzing data is actually the easy part?

At this point, we’ve all probably heard the classic pitch. BI is hard because of siloes — analysts on the data team handle the legwork, but business leaders own the context. Tableau licenses are expensive, so view-only seats it is. Executives are still working in spreadsheets while self-driving cars hit the gas pedal all the way up hills we’re still scared to drive up in San Francisco.

Of course, there’s a lot of nuance here that’s contributed to this age-old problem snowballing into a more pervasive one. For one, manual work aside, Excel is a perfectly valid solution that gets the job done in the early stages of a company’s lifecycle.

Analysts and execs also fundamentally don’t approach getting answers in the same way. Benn Stancil, cofounder of Mode, puts it well in his article — analysts “work like scientists, creating new datasets and aggregating them in novel ways to draw conclusions about specific, nuanced hypotheses. Non-analysts work like journalists, collating existing metrics and drawing conclusions by considering them in their totality.” Incorrect filters, constantly shifting product pricing/positioning/everything else, and technical metric complexity add to this divide and breed mistrust over time.

It’s a real problem that affects almost every business today - one that a lot of very, very cool new companies are trying to solve in extremely interesting and bold new ways.

And yet, in conversations about data-driven decision-making with growth and revenue leaders that work at a large range of software companies, frustration almost always seems to be directed at something slightly different.

We’re splitting these sentiments into three overarching categories.

1. For very technical and tech-literate teams, understanding data tends to usually be pretty easy. Everyone already has a gut feeling and their data team enables them to work decently quickly so they’re relatively happy with their processes. It’s the last mile from insight to action that’s almost impossible. End users get analysis paralysis when metrics are too complex, dashboards aren’t easily understandable, or when there isn’t an easy way to distill new information down into a few paths of action that they or management can then decide from.

2. Enterprise orgs that tend to have a large team of sales reps have well-resourced data teams, but revenue operations and analytics teams naturally handle a lot of the demand for answers from sales, marketing and customer success. Many of these teams are mainly looking for visibility in the form of a scoreboard everyone can see that tracks the health of the overall business. They want a level of flexibility Salesforce dashboards and reports don’t currently support so they can create “run your business” dashboard. Being able to do proactive analysis and push forward recommendations instead of only understanding what’s happened historically.

3. And of course, the seed-stage(ish) startup. They’re in massive experimentation mode, really just trying to figure out what’s going to work. One or two people might be in charge of all things data/writing SQL/joining tables/making pretty graphs, and often times what they really want is a to plug things in a “what if” machine as efficiently as possible and be able to measure the result. Things change fast, and most of the time they don’t have any real historical data to benchmark off of. Getting metrics does take a lot of manual work, but their bigger challenge is figuring out which signals are most pressing and what to do about it.

Don’t get us wrong — better BI is a very real and important win for everyone. But from what it sounds like, it’s only the foundation to unlock true, data-driven decision making (yet another buzz word I couldn’t think of a good synonym was) — the real challenge lies in building the rest of the house.

We kind of get the gut feeling that at this point you might be thinking about how AI fits into all of this — or maybe you’re not in which case we feel like the last sentence was a pretty good closer and is a good place to stop reading.

If you’re sticking with us (we’re flattered), one line of thinking that’s pretty interesting is the idea that we may have gotten our role in the creative hierarchy all wrong. We clearly really like Benn’s blog, where he codifies that “one of the most striking things about LLMs is that they’re much better at the creative parts of analysis than they are at the mechanical parts”.

A look back into our most “successful” recent queries — where to get dinner in Cape Cod, best beaches in Cape Cod, what to do in Cape Cod over the summer (yes, one of us went to Cape Cod), how to condense this article into a LinkedIn post with ten too many emojis, synonyms for the word “transparent” — seem to support that idea at first glance, since we still pull out calculators to do routine math and code APIs relatively by hand.

Is it that far of a jump to use LLMs as a springboard for the much more high-stakes, complex decisions we face at work where we feel similarly, if not more, creatively stalled? Half the synonym results for “transparent” were irrelevant or incoherent, but there’s something about being able to sift through someone else’s ideas, no matter how off, that opens your own mind up. We might be a few no-fluff, grammatically incorrect commands away from a breakthrough.