I joined RJMetrics in 2013. We were building a business intelligence product primarily for ecommerce companies. We believed that businesses could run more efficiently with data, and anyone — not just analysts — should be able to get their hands on that data as needed.
As we started to expand into markets beyond just ecommerce, it became clear just how hard this was. Datasets were messy, every business had custom needs, scale became difficult.
The marketing team was grinding on lead gen, and we were actually pretty good at it, but growth felt really, really hard.
Then Redshift came along, and almost overnight, the business intelligence landscape transformed. With access to an unbelievably powerful data warehouse optimized for analytics, market demand shifted from “full-stack analytics software” to “something-to-help-me-get-all-my-data-into-redshift-right-now-please!”
We pivoted, building a product initially called Pipeline that did just that — it got data into Redshift. It was a wild few months culminating with Magento acquiring the RJMetrics classic solution (a perfect ecommerce fit), and then launching Pipeline into a whole new company called Stitch Data (recently acquired by Talend for $60 million).
I stayed on through the acquisition and helped launch the Stitch brand and then moved on…
I was a HubSpot user back in 2012 and had been a firm believer in the inbound methodology, but my time at RJMetrics made it abundantly clear to me that great content marketing alone wasn’t enough to grow a business. I wrote this back in 2016 when I started working for myself for a bit:
“So here I am, a marketer whose whole career is rooted in content and I am deeply suspicious of the passionate preachers of inbound. It’s not that it doesn’t work, I still think content marketing can be incredibly impactful, but I don’t think my experiences are an anomaly. Inbound marketing still works, just not quite like everyone says it does.”
But then “the passionate preachers of inbound” read my post and got in touch with me about working for them. I was a little surprised. During my interview process we talked about how inbound was changing. I was blown away by the honesty, brilliance, and humility of that team (still am).
During my time there, these ideas around inbound changing and the challenges of customer acquisition began coalescing around what our CEO, Brian Halligan, called “the shift from funnel to flywheel.” The funnel is this idea that businesses invest in the top of the funnel and customers come shooting out the bottom. It was an important concept in the early days of digital marketing when we were still trying to figure out how the math worked, but today measurement tools are more sophisticated, acquisition channels are fragmented, and customer retention is recognized as being as important as customer acquisition. The funnel is no longer all that helpful as a concept to describe growth.
HubSpot calls this new concept the flywheel — which nicely describes the idea that happy users talk about your brand, acting as an acquisition channel, and then you add more users who become happy and vocal themselves, and momentum builds.
Everything about this made sense to me.
Our marketing strategy at RJMetrics was entirely focused on throwing time and money into the top of the funnel — building advocacy wasn’t what marketing did or cared about. Sure, we put together some great case studies, but we didn’t own advocacy in any meaningful way.
Throwing money and time into the top of the funnel before you’ve figured out customer retention and advocacy is exactly where our marketing strategy at RJMetrics went wrong. It’s what I was trying to describe when I wrote Fix Your Marketing Engine, Then Add Gasoline.
Fishtown Analytics and dbt
While I was doing my thing at HubSpot, a few of my former colleagues started Fishtown Analytics and the open-source software that powers Fishtown Analytics work, dbt (data build tool). dbt is a response to doing analytics in the world of Redshift and is built on a fundamental assumption that earlier business intelligence tools didn’t have the luxury of assuming, which was this — analysts should work like software engineers, mainly, they should write code rather than interact with fancy data visualization tools.
Earlier BI tools couldn’t assume this because there was no Redshift. So analysts and engineers writing code would hit your production database. It was risky with the potential to take down your site or slow order processing. With Redshift, that’s no longer the case. Data is extracted from production databases, loaded into Redshift using tools like Stitch, and then analysts can conduct transformations with no fear of messing up critical business systems.
dbt handles those transformations or “builds”.
This sounds small. It is not.
dbt represents a completely different framework for thinking about how to do analytics. It assumes that the work of analyst is not just “build charts” or “build reports” but is instead “model data” and “create trustworthy data sets.” dbt assumes that analyst can and should scale themselves through code.
With dbt, analysts can do the work of data engineers — handling data directly rather than relying on an intermediary. Analysts can get off the analysis treadmill of creating more charts and furiously maintaining accuracy, and can become data stewards.
Not surprisingly, analysts who make the fundamental shift to working like an engineer, become hooked. They get promotions. Their teams grow. And most importantly, they become super-vocal advocates of dbt.
The flywheel is cranking for dbt. So when the team asked me to join them as marketer #1, I jumped. To start, I’ll largely be helping to meet the demand of users who want to talk about what they’re doing with dbt. They want their logos on the Fishtown Analytics site, they want editing support when writing articles like this about how they are using dbt, they are asking to do case studies.
Everything about this feels right to me. The role of marketing in this environment is no longer “creating demand” for a product, but rather meeting the demand of users who want to do our marketing for us.
That’s pretty cool.
I’m looking forward to being part of a Philadelphia startup again, and one with so many crazy enthusiastic users. I’m looking forward to being back in the rapidly evolving world of data analytics. I’m looking forward to working with a team of people that I admire so much, some of whom I’ve worked with before, others new.
Huge thanks to HubSpot for a remarkable 2+ years. I cannot recommend the team there highly enough, and I am honored to have had the experience to work with so many smart, humble, talented people.
Here’s to the work ahead!