After four years at Pinecone, I'm back to consulting

After helping Pinecone go from an unknown pre-seed company to a household name in the AI space, I'm returning to independent consulting (and writing).

Mission-driven

In 2013, the same year I started consulting, I got the idea to run the New York City marathon. That idea turned into an irrational and stubborn mission.

That mission motivated me through training runs in all kinds of moods and weather and places and physical conditions. And when I didn't win the lottery to secure a place in the race, the mission kept me motivated to run other marathons for three consecutive years, always with an eye toward the NYC event. Running became part of my identity and directly impacted everything from my schedule to wardrobe to the music I listened to.

Finally in my fourth year of running I secured a spot in the NYC marathon. On the rainy afternoon of November 5, 2017, I crossed the finish line and my mission was complete.

And then I stopped running. Not because of an injury, or lack of time, or anything else like that. Simply because in the absence of a mission there was nothing driving me forward — or rather outside to run in circles.

The mission at Pinecone

In 2020, I got the idea to join Pinecone. Actually Edo Liberty had the idea first, after we spent three months on a consulting project. It was just one of several crazy ideas we had, along with:

• Marketing an AI product to developers. At the time, vector search was a niche capability utilized by a relatively small number of data scientists and expert search engineers using one of several open source libraries. There was a growing number of "ML engineers" but it would be another two years until ChatGPT and developers en masse started building AI applications.
• Creating a new category and market which we called "vector databases." I was well aware of how much of a long shot this was; just months earlier I warned that category creation is much harder than most people expect.
• Taking a product-led approach to growth (PLG) with a fully managed product (ie, not open-sourced). Conventional wisdom at the time was you had to open-source to have any chance at building a community and winning over users.

This was startup marketing on ultra-hard mode. I knew it would be a grueling and long-shot marathon, and I knew by then I'd need a mission to power through it. Within the first month I had one:

The mission of the marketing team at Pinecone was to grow adoption by helping engineers discover, explore, and build with vector search.

This mission remained unchanged and underpinned our marketing strategy and all our execution for the next four years. By the end of which:

• The vector database category is considered a core piece of AI infrastructure, with Pinecone still the first name in that category.
• We had $138M in funding from top-tier firms like A16z, Menlo, ICONIQ, and Wing.
• We reached 6,000+ paying customers, the vast majority of them coming through non-paid (organic) marketing channels and into the self-serve product.
• Hundreds of thousands of developers have used Pinecone and countless more have heard of it.

As much as those accomplishments filled me with pride and gratitude for the opportunity, in the past year they increasingly made me feel adrift and without a purpose. Because my mission of getting developers to know about and value vector search was more or less complete, and my work started to feel less like cathedral-building and more like laying bricks. Or, like running in circles.

By August of this year it became clear to me and Edo — by then a close friend of four years — it was time for me to pass the torch and find myself a new mission. And so I did.

The Next Mission

The most impactful software products often start as deeply technical solutions that few people understand, aimed at technical audiences like software engineers and data scientists that few people know how to market to. So the ultimate success of the product often hinges on whether you can get those technical audiences to find, understand, and value the product enough to adopt it.

That's... extremely hard to do, even with a phenomenal product. As it happens, I'm exceptionally and demonstrably good at it, and I love doing it.

That's why I'm returning to independent consulting with the mission of helping companies get deeply technical audiences to understand and value their deeply technical products. Put simply: I help companies win technical audiences when it's crucial for their success. (The more technical the product and audience, the better.)

Starting today I am doing this in two ways:

1. Taking on audacious consulting projects (2–6 months) with software/AI companies, in partnership with the founding or executive team. I'm delighted to already be working with several companies and expect more availability in April of 2025. (If you could use my help, send me an email: greg@gkogan.co.)
2. Publishing more content on this blog — including lessons not only from my time growing Pinecone but from the 40+ startups I consulted before that. This way I hope to reach and indirectly help many more founders and teams than I can possibly work with one-on-one. (If you want to get those posts sent to your inbox, you can subscribe for updates.)

Gratitude

I'm tremendously grateful to Edo, the entire Pinecone team, the investors and advisors, and the whole Pinecone community for the past four years. It was the run of a lifetime and, for me and for the first era of Pinecone's long journey: a mission complete.