We spoke about Bluecore’s revenue forecasting app, how the team prioritizes projects and measures success, reasons for adopting Hex, and more. The interview has been condensed and edited for clarity.
If you missed Part 1, where we introduced Bluecore’s data team and modern data stack, use cases for Hex, and other topics, check it out here.
Alex Saccois Bluecore’s Senior Director of DataOps & Analytics; before Bluecore, Alex held Analytics, BI and Strategy roles at AlphaSense, CB Insights, and Luminary Labs. Adam Whitaker is the Go-to-Market Analytics Lead for Bluecore, with previous Data & Analytics roles at Gopuff, Monetate, and Boeing.
Adam: I started in September and we delivered it the first week of January. So three months minus holidays and my ramp time. And there was a lot more than just building out the Hex content itself. There was a good amount of dbt work that needed to happen ahead of time in order to stage up the data.
We needed deep benchmarking. We needed to be confident - and help our clients be confident - in a few things.
1What baseline performance should they expect? What will your next year look like with Bluecore?
How could you improve from that baseline? What opportunities do we have to drive better results, and how can we prioritize those?
Solving for these helped us understand and explain performance both across our business and on a client-by-client basis.
Adam: We originally built it for our Existing Business team, who was going to use this tool to manage client renewals. They could see the client’s historical results alongside data-driven forecasts for future performance. That supported conversations around renewals.
One of the unexpected outcomes was that our internal Customer Success (CS) team started to pick it up and use it to find optimizations for clients. They’d ask questions like:
We've got 20 different email capabilities, my client is using five of them. What if they turned on number six? What would that look like?
Here's a specific use case that I want to recommend to the client. What type of performance could they get out of that?
If the client tells me they’ll implement one recommendation, which strategy will have the biggest impact on the business that they could not do before?
So it turned into both a CS tool for managing the account week to week, month to month. That’s on top of the Sales use case of projecting consumption and results to structure the renewal contract.
Of course, there was a learning curve and some pushback. A lot of people in Sales don't want to use tools like this. We've been challenged a number of times from these folks and we've been able to reply back with some killer onboarding and examples, and the outcomes are aligned with what those sales reps are expecting.
People who were total doubters have said, “You know, this saves me a ton of time.” That’s where the benefit of focused documentation and enablement really shows through.
Alex It's an exciting challenge for any centralized data team because you're a shared resource across the org. We find ourselves in the fortunate circumstance of shipping work across the exec team, e.g., SVP of Product, SVP of Engineering, CCO (Chief Customer Officer), CFO, VP of Finance –and to some of our largest clients. There are three key conventions that help us navigate this:
1. Org design. We’ve leaned into a domain or federated structure. This creates deeper business knowledge, better relationships, and lower context switching. It also means capacity tradeoffs can be made “locally” within a given department, rather than across the whole team.
2. Roadmapping. dbt famously advocates for “applying the best practices of software engineering to analytics.” We couldn’t agree more! That’s equally true for capacity planning and roadmaps, especially for a central data org. We actively participate in org-wide planning efforts, working closely with our eng org to call out dependencies, and are accountable during org-wide status reviews. This not only gives the DataOps team more clarity on goals, but also helps other departments learn about our process, footprint, and impact.
3. Communication. Pod or federated data structures work well for velocity but risk creating isolated work and missing the big picture. Internally we have regular demo days to share learning and best practices. Externally, we have regular prio review meetings with C-suite execs, and send written updates to track our progress against goals.
Above all, having executive leadership that is bought into the investment in data as a core competency has been a huge tailwind to our team’s success. It’s enabled us to hire great people and balance long-term strategic work with quick wins.
Alex: We wanted a tool as powerful and flexible as Jupyter notebooks, as shareable and collaborative as Google Sheets, and as scalable as an enterprise BI platform.
It’s a little on-the-nose, but that’s how I view Hex. We can create powerful reporting and analysis that’s fully shareable and replicable. We can prototype transforms and even UI analytics features at incredible velocity, with potential for deep collaboration. There’s even a customer experience win, where live apps can be shared with clients to answer key analytics questions.
Adam: At Bluecore, we were all using Jupyter notebooks to begin with. They were all managed locally and they were totally not shareable, which is bad. We now have a shareable solution that is a better Jupyter notebook for sure. That's the baseline.
We have something that we don't have to host. That’s awesome. All of the packages are pre-installed for us, we don't have to worry about that stuff – fantastic.
I can share and we can collaborate at the same time. In addition to that, we don't have to be pivoting between creating data and then figuring out how to share it and post it and all those things. There's interactive capabilities where we could build a dropdown that allows somebody to select something and then save that thing off or fire off an API call, and then on top of that, we can build reporting. I don't know how many times we’ve built a Looker report and then somebody's like, “That chart just needs to have this thing different.”
There's just that level of customization that doesn't exist elsewhere. We have full blown Python packages that we can install and then customize ourselves, and we can really create anything that we need to at that point. It's a huge improvement over the workflow that other tools offer.
And then the integrations that are happening; up to now, there's been the dbt integration. All of the Jinja stuff is great. All of the scripting and looping and controls are a big bonus for us, especially when we're building these apps that have deep customization inside of them.
And the scalability, repeatability. I haven't had to go back through my old notes and emails to find a query that I shared with somebody or a file that was sent out to figure out how I did it. Everything is in Hex. I literally use it for all of my stuff these days. I'm a big fan. I appreciate the tool in a big way.
Adam: I was content with Jupyter notebooks and didn't know that there was something better out there, but got my hands on Hex and said, “Wow, this fills a huge hole that I didn't know that I had, but now I do. And I'm grateful to have it.”
Alex: From a team point of view, it’s like the kitchen at the Gramercy Tavern or the service bay at a Porsche dealership. If you want to bring in the best people who are excited to ship amazing work, you give them the best equipment.
I want to have best-in-class tooling because it's a very worthwhile contribution to the productivity and happiness of our team. It just lowers the friction of delivering work in our team, and across the business as we scale and enable new users with tools like Hex.
When we were shopping for tooling, I wanted a more flexible environment that was going to be more collaborative. That was really compelling with Hex. The other part is the client shareability – that was a decision-making attribute because we needed a way outside of Looker to create live, one-off, client-shareable reports that wouldn't be beholden to BI layer constraints. We needed modular, granular shareability.
That Google Sheets-eque shareability of Hex was key because we knew we needed to meet these client data requests in a more scalable way, somewhere between full, in-UI, universal for all clients, and a spreadsheet. And Hex nailed that.
What has proven to be really interesting is how much of a Swiss army knife Hex is. We're now using it to ingest data from several one-off APIs and make custom reporting. Six months ago, that would've been a multi-team, huge effort. Now, a less technical member of our team is able to execute that in a couple of sprints. And the client team reacts with comments like, “Oh my God, we expected this to take a year!” The velocity that Hex has unlocked has been really amazing, even against high expectations.
Adam: There's a lot of scaling, testing that we are going through right now. There's a huge eye towards making sure our technology stack will scale to the size that we need. How do we make sure that we don't have a fire on those four to five days of the year?
Alex: It’s a critical time for our clients, and our goal is to give them flawless performance that helps them drive extraordinary results.
And one of the ways Hex is going to be playing a role is that we do a lot of very fast cycle analytics and reporting about BFCM [Black Friday Cyber Monday] because Bluecore's BFCM benchmarks and metrics report is popular in the industry. So we want to be able to flexibly and scalably create these EDAs and other reports for data marketing near real-time after BFCM. And when we have something like Hex, the scalability, replicability of that is best in class.
Adam: Instead of us being awake at whatever time of night or whatever time in the morning, we're going to head into the holiday building out views in Hex and testing the heck out of them. And then we give that tooling to a Marketing team. “Here you go. If you want to work at 1:00 AM, go for it. We'll see you the next day.” We’re giving people the tooling that they need to be able to do their jobs without us having to be on call at all times.
Alex: That is kind of a story with Hex that is underappreciated – the whole idea of building for some future state, so if the Marketing team, the Ops team want to hire more technical people under an embedded data model in next year, the learning curve for them is much shorter because we have this installed base of very transparent, best practices tools.
Learn more about Bluecore’s forecasting app – catch Adam’s Coalesce session on demand: Data automation with dollars on the line: Forecasting 7-figure deals with Hex, dbt and Hightouch.
Catch up on Part 1 of our Q&A with Bluecore to learn more about Bluecore’s data team and modern data stack, use cases for Hex, and more.