Good GTM data is hard to find
I've only seen one GTM data environment that really got it right. My first audit allowed partial credit, which hid the real gaps, so now it's 1 or 0.
When I started in go-to-market consulting, I did not fully appreciate the variety and variability of data I would find moving client to client. Everything is a mess, everywhere and as long as the company is making money, most people don’t really care. Now, when the money stops, or when growth doesn’t meet expectations, a host of consultants might be summoned to look at what’s wrong and to make important strategic recommendations from absolutely disastrous data. I assume at all times that everyone is doing the best they can with their resources, but it’s usually not a problem that can just be cleaned up in 2 weeks or 2 months. It’s also not something you can point Claude or ChatGPT at and tell them to fix because GTM data is an output, reflection, or projection of the systems and processes that create it.
My last client before re-entering the corporate arena actually had the most pristine data environment I’ve ever seen. Every function had a data leader sitting on a council to review changes to the data lake every two weeks, and there were several important data quality directives imposed by the engineering lead, who I still consider to be the best in the world at this discipline, that have shaped my views on what good looks like over the last decade. They tell you if you find Utopia, just keep walking because you’ll probably ruin it. I didn’t listen, I got permission to take a role there from the partners, had 3 months of rolling around like a pig in the mud in data Nirvana, and then was promptly acquired by Big Tech, which is a whole separate topic for another day and a few beverages.
When I decided to go back into the 100M - 1Bn range of tech, I did so with a new found conviction and perspective on GTM data with an urgency driven by the arrival of limitless possibilities with frontier AI models. The stark disconnect between what people knew should be possible with such a generational technology and the early returns companies were actually getting is eye-opening. Especially confounding since my hypothesis is that the anatomy of B2B SaaS and to an extent B2B anything is universal: prospects, customers, contacts, deals, subscription contracts, invoicing, and usage.
I ran my first two weeks at a new stop like the audit that became the building blocks of Foundat10n. I started with a few simple questions about the source of truth for board reporting, how snapshots were being maintained, and what the prevalent opinion on these outputs were after being retrieved or used by their AI of choice.
That first audit actually reinforced a stricter view shifting from a spectrum of maturity to a pass/fail approach. Second order and third order sets of analyses started to attenuate exponentially when pieces of this were “partially” implemented. So while the 1/0 framing might be harsh, and convenient to the wordmark Foundat10n, it is not without merit.
Like anything, there’s not just 10 things you have to do to be great at something, but I picked these 10 because I know you can get all of these off the ground in 3 months. I’ve done it with my own hands with only the assistance of Claude’s many hands. In-so-doing, stakeholders start to actually feel how much better their AI outcomes are with this approach. Trust in AI outputs increases because the numbers match the dashboards and reports. When the answers are read from a single surface, the agents become more consistent. And strategically, your teams start to get a better sense for what they should own themselves and what they should buy or rent from others. An early success creates a flywheel that can catch the rest of a company’s data ecosystem in its momentum.