Everything is hard at a startup. But everything is also an opportunity. This is a behind-the-scenes look at how good ideas win at Sortable. It’s messy, divergent and a good reflection of what it’s like to work at a tech scaleup!
I’m Daniel - I’m a software developer and team lead for Ad Reports. I’ve been with Sortable since May 2016. Before Sortable, I was doing my master’s in Paris, and interning at SAP.
Right now, a new report system that I helped create is in beta. Internally code-named The Dream for its ambitious product goals, the new reporting system has a 100% adoption by internal staff. A couple customers have early access. By mid-summer, it will be available to every Sortable customer. Given that everyone thought it was impossible, these are considerable successes.
This post is about how we started The Dream.
One year ago, there was a need to expose data in our data warehouse to some of our customers. The simplest solution was to build a system where our internal analysts could write custom queries and share the results with users via our reporting portal. Later, the users would be able to refresh them so they could always have the latest data. We called this system Custom Reports. Custom Reports would give our users access to more granular data and use the visualization system that we created for the previous modules, but with the caveat that our users were unable able to edit the reports.
Partner Report - The predecessor to The Dream
Close to finishing Custom Reports, we received a requirement from one of our internal teams for easy access to our Demand Partners bidding data. This was something else that we stored in our Data Warehouse, but didn’t expose in our reporting UI.
So we decided to create a system that allowed users to interactively report on this specific data. This project was developed by our co-op David. This project worked really well. We solved the Demand team’s problem and also demonstrated that it was possible to have a system that allowed dynamic queries to our Data Warehouse.
In the weeds
Partner Report was so useful that we started getting requests to expose it to customers. The tool wasn’t ready for that since it was a co-op project that needed to be refined.
But we identified a major opportunity from this request. We had Custom Reports, a window where customers could see a part of their data but couldn’t touch it, or play with it. We needed a door into the data warehouse.
Proof of concept
That fall, we got another request to interactively expose a new part of the Data Warehouse.. The task was not well specified, so when we took it on, we did it in our own way.
We had an idea to make something that was more dynamic and flexible, that could scale with our needs. We knew that solving everyone’s problems by writing custom, one-off reports would never scale. We decided to build a new tool that had the ability to report on our entire data warehouse. Because all of our tables have a similar structure, if we solved the problem for one table, we basically solved it for all of them, which could be a huge win.
During our sprint planning, when I told the team about our idea, most of them were against building a separate screen, because it would take more time. But team lead privileges meant I could work through people’s concerns by committing to a proof of concept.
We had two weeks in the sprint to prototype. If it took any longer, then they would be right to tell us to shut it down and do it with a Custom Report.
Sidenote: An ode to Jordan
Jordan is a software developer on my team and does most of the UX. He built most of The Dream backend and frontend. I got Jordan inspired to jump onboard, but he had one condition. He was concerned that it would get shut down if it didn’t work. He agreed to help if I sold the prototype to Bryan, our VP of Engineering.
In less than two weeks, Jordan managed to build an ugly prototype of The Dream, reusing parts from all over the place. People could tell us we were crazy when it was just an idea. But when they could see it, they realized it wasn’t that crazy and it could work.
I brought in Bryan and introduced him to The Dream. Our data warehouse is complex and few people will understand it. So the pitch to Bryan was: if you have a few people who understand it building a query tool on top of it, nobody else needs to worry about the SQL. You click on the thing. This interface would solve internal, external and future problems.
We built those two new reports in a week as proof of concept. Bryan was happy with it and approved it to go forward. Next, I pitched it to Chris, our CEO. He liked it. We went from prototype to let’s make it happen with key stakeholders. The idea of the Dream was now a reality.
The Dream goes live at Sortable
When we had something a bit more stable, we introduced it to the business side of the organization at the start of the new year. We scheduled demos of new features and received wonderful feedback about what was useful, what was broken, what could be improved. Today, everyone in the company has full access to The Dream, and they’ve given even more iteratively useful feedback. We’ve moved out of the proof of concept phase and into a constant feature for customers.
So that’s how an idea became a feature at Sortable. The lesson is that rapid and continuous improvement is critical for success. The best ideas win when you try something small, fail fast, discard what doesn’t work, and iterate the idea with people. You repeat that until you can sell the company stakeholders on the vision that’s been tested and won.