"Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Brian Dolan, founder and president at WorkReduce.
I was in a bar with an ad tech CEO who was describing a visit to a luxury brand CMO.
“It’s my first trip with this sales rep, and the meeting’s going well,” he said. “Then the CMO asks, ‘What’s your one key differentiator?’”
He paused for emphasis. “And the rep says, ‘Our algorithm.’”
“Suddenly we’re so deep in the weeds I can’t see daylight,” he said. “I saved it, but it was close, and I can’t afford to ax the rep. Decent sellers are just too hard to find. What should I do?”
After all, the dictionary definition of algorithm is as bland as an off-white wall, and so are buyers’ reactions: it’s a set of instructions for carrying out a process. A recipe.
So you have a recipe. Whoopee. No sale.
Yet in today’s tech-heavy, big data-driven, algorithmic advertising world, you can’t pitch a tech play without talking about the technology itself. It’s a trap: Even if your algorithm truly is a better mousetrap, bringing it up is a route into esoteric territory with a nontechnical customer. What’s the way out?
In software, it’s all about algorithms. When you study computer science and software engineering, the first and most fundamental algorithms you learn are for sorting. And it’s amazing how often things need sorting in software.
And it turns out there are a lot of ways to sort. Your algorithm choice matters. Performance varies widely based on the number of things to sort, a list’s initial state and computing resources at hand, such as storage, processor and memory.
The difference is obvious when you visualize how they work. Check out animations of six common sorting algorithms here (click the green arrows to start the animations).
Without an animation, how can you tell the difference? Math. Enough for a dedicated undergrad computer science course. Or you can run the code.
But absent math or code, the difference between sorting algorithms is inscrutable.
Software engineers buy software, too. Nerdy software, like databases. Everyone needs one at some point. No one wants to write one from scratch. And databases have algorithms galore.
But how are databases sold? It’s simple. Based on performance and cost: number of records handled, support, maintenance, hardware required, integration, license models and a host of other factors.
Even when geeks buy software, algorithms only come up as an afterthought or in passing. No one looks at the code.
So you can’t be an effective software engineer without understanding how algorithms work, yet as an engineer, buying ultra-geeky software, you’re probably not talking algorithms.
So how do ad tech sellers avoid the same trap?
Interpretation. Like the visualization of sorting and the performance/cost characteristics of databases, in ad tech it’s the outcomes and benefits that matter, not the algorithm.
Advances in ad tech have delivered increased reach, better cost efficiency and new ways to tell stories to consumers. And we’re just getting started. But along with the good comes the bad: confusing acronyms, a cluttered vendor landscape and traps like “It’s our algorithm.”
The deep irony of selling ad tech is that what it needs most is you, a human.
Only people – ad tech sellers and their support organizations -- can provide the layer of interpretation buyers need: the link between technology and tangible benefits. Tell a story about results. Explain the costs and risks as well as the upside.
The next time you’re in an ad tech sales pitch and hear the word “algorithm,” check yourself. Something important is missing in action.