Businesses operating today have a complexity problem, and it's paralyzing data teams. That problem stems from many roots: growth demand from investors is higher than it has ever been; it costs more than ever to start up; companies have larger potential audiences with a longer tail of interest; nearly all processes are measurable and trackable; and vendor choices have multiplied.
True, there are large companies that have cornered markets and need real sophistication to keep generating new customers and upselling them. That said, if your company isn't among the Apples, Amazons and Walmarts of the world, you likely don't require the same machinery under the hood that they do to run a profitable business. Customers have historically made purchase decisions the same way they do today: based on price, value, need and convenience. Beyond these, there are social-emotional triggers for purchase (mood, a solid recommendation, aspirational lifestyle), of course. Whoever said customers were fully rational has never met one.
The above levers are tough to pull at a moment's notice. Businesses have costs to produce products, plus an immediate ceiling on the quality and convenience they can offer. Moreover, these base principles are the toughest part of a purchase to measure. Most consumers don't tell you why they bought something, or why they bought it from your competitor instead of you. It can even change from purchase to purchase. People are far more likely to leave negative feedback than positive and most won't leave any at all. And thus, we've arrived at part of the complexity: how do you drive future sales when you don't have full information on past sales?
The answer has been proxies, and proxies are a tough data problem.
I won't go into the many possible metrics one might use to optimize business performance, but a quick Google search or Claude query will make you dizzy. Some businesses can run well on metrics that yours can't, and vice versa. It requires heavy costs, infrastructure and buy-in to run well-executed experiments in a business setting, and the value isn't always apparent or long-lived. Measurement itself can be expensive, since it often requires software development or a product in its own right. So what is a business to do?
Go back to the basics, I say.
Most businesses make their money from a small percentage of SKUs/products and a small percentage of customers (I'm looking at you, Nvidia). Many businesses with high rates of repeat customers tend to be profitable, and those without tend not to be. These are your north stars in a complex business world: margins, and a quality product solving a real, repeated problem. These metrics aren't the most interesting data problems to solve, but they back a sound business. If your C-suite can point to their top products and how they perform on units and revenue over time, the health of their relationships with top clients, and whether the business is hurting or helping itself while trying to acquire new customers and develop products, it's much easier to make good decisions on your core model and your long shots. And if you can't do that, what good is it to optimize against the friction of a dropped cart or a site click from a marketing campaign, anyway?
I realize this oversimplifies the subject matter, but it's to make a point: simple is usually the path best taken. It's often the most cost-effective and efficient use of time. It tells you 80% of the answer with 20% of the work. That's the whole philosophy we build around at Data Culture: we're the team that helps you find and build the 20% that matters instead of overengineering the rest.
Most of the time, if the underlying data that answers fundamental questions is good, it'll give your business the solid foundation it needs if a more sophisticated solution is ever required. When other businesses are building expensive automations in niche corners of their data warehouse, I will be shouting from the roof top to pare down dashboards, find source of truth columns, and document your data. And if you can't hear me, try to remember why there are so many laundromats and coffee shops.
