dbt has an increasingly-large ecosystem of service providers: from smaller analytics-focused shops all the way up to some of the largest consulting organizations in the world. One of the questions that we, Fishtown Analytics, get from these companies is related to channel competition. I wanted to take a minute to talk about this topic in a public forum; my hope is that by saying some of this stuff publicly it will make our stance and commitment on this topic clear and unambiguous.
The context here is that Fishtown Analytics, for a very long time, supported itself completely by professional services. This was essentially true from July of 2016 through November of 2019 when we raised our seed. During this time we helped companies both implement and do analytics work inside of the modern data stack. Every one of our projects had dbt at its core, and this enabled us to both make money from and iterate on a piece of open source technology that had no other revenue streams at the time. There are a ton of good things that have come from this bootstrap-open-source-tech-using-services approach and I highly recommend it to other folks starting an enterprise software business!
BUT! Over time, as the dbt community has grown and other service providers have been interested in also participating in the ecosystem, there has been this ongoing concern from these service providers that they’re in competition with Fishtown Analytics for services revenue. This may have been a completely legit concern back in mid-2019, but it’s not really reflective of how we think about our business any more. Here’s some more information that’ll help you see why that’s true.
- We’ve made the commitment to be a software company. We now have some very A-list investors, and these folks didn’t put money into us to be a professional services company. For those of you who may be less familiar with the ecosystem, VCs don’t really invest in professional services companies. This is for a ton of reasons, the main two being a) scaling professional services is really hard, and b) the gross margins for professional services tend to be ~30% vs. software’s ~80%. This isn’t the kind of decision that we can just “go back on”…we’re committed to this path.
- We’ve already shrunk our professional services team. Depending on how you counted it, we had ~8-10 humans working roughly full-time on professional services in October of 2019. Today, that is more like 5 humans devoting roughly half of their time to professional services engagements. We used to run about 20-22 simultaneous client projects; today that number looks much more like 8-10. (For reference, our projects average ~10-15k / month.)
- We’re focused on services that are strategic for our software business. An increasing percentage of our professional services team’s time is spent on things like teaching dbt Learn and acting as internal dbt subject matter experts vs. core “consulting” projects.
- We directly partner with other service providers on projects. In our largest dbt Cloud Enterprise clients, our professional services team is typically supporting a much larger consulting engagement that is being run by a partner. Rather than attempting to deliver these projects ourselves (and earn the very large consulting fees that they entail), our goal is to “sprinkle in” a little bit of our time to make sure that the larger project feels supported by Fishtown Analytics and the deep product knowledge that we can bring.
Hopefully, the above makes it pretty obvious that we have absolutely zero interest in attempting to capture the large amount of professional services work that will happen in the dbt ecosystem. I’ll also say that, personally, I’m just not that obsessed with the question of how to scale a consulting organization. Having scaled one to ten people, it’s really freaking hard! And it’s not the kind of hard that I’m particularly good at or interested in...which is why it’s great that there are so many other humans who are wired differently and are much more suited to solve it!!
To be clear, we will continue to do some professional services work forever. This is for three very specific reasons:
- All enterprise software businesses provide some level of professional services support in-house because enterprise customers demand it. If we didn’t have any professional services competency internally it would significantly constrain our ability to sell our software. Looker provides this, so does Snowflake. It shouldn’t be surprising that we would do this as well.
- Professional services work is a bedrock of our company culture. We like to say “we do the work.” This is at the very heart of what has enabled us to build dbt—we have primarily built a product that we found useful ourselves! We’ve always approached product design from that perspective, and if we all the sudden stopped doing any analytics work with dbt in real life settings then we would also lose the insight that has enabled us to make the product great in the first place!
- Having a professional services practice allows us to have a fantastic internal talent pipeline—some of the folks who do our professional services work end up taking on roles in other parts of the company and have a massive impact on our product and community. Without this ability to create real-life experiences for these folks, they wouldn’t be able to create the kind of impact that they now do.
To be extremely, painfully clear: none of those reasons is “we want to make a shitload of money on this.” Professional services revenue is just not at all the point. And while we will likely hire more folks on our proserv team in the future, in no way is our goal to scale this part of our business in a big way.
Here’s my guess as to how this will play out. The dbt ecosystem is growing quickly, and service providers in the space will see more and more revenue generated from it. We’ll have more opportunities to work together on deals that include some software revenue (for us) and lots of services revenue (for you). As this happens more and more, this question of channel competition will just fade into the background—an artifact of a past era that no longer feels that relevant. The total size of the opportunity here is far bigger than what any of us sees today. That’s what we’re building towards as a company and as a product, and we’re excited about partnering with you to build towards that as well.
If you have comments / questions on any aspect of this, please feel free to shoot me a DM in Slack!