Tandem Version 1
Pretty shortly after we started Tandem we spun up a Git repository named Horoscope. It contained our company handbook and the things we wanted to do to build out Tandem as an organization. We'd get together every Friday and talk about problems surfaced in recent client work and how we could improve them. Then we'd set some priorities, assign some tickets and in the downtime between client projects work to improve Tandem as a business.
In the beginning this worked great! We quickly documented our company mission, values, and important processes and procedures. We crafted an employee handbook. We developed some engineering and workflow standards and automated a good deal of our DevOps. However, as is often the case in #agencylyfe, we increasingly had less and less time to dedicate to this cycle of iterative improvement. We became complacent and disorganized. And while we all, to our credit, continued to build valuable things we struggled to incorporate their value back into the "bigger picture". In some cases we even failed to communicate their existence altogether, leading to a lot of duplicate work and unnecessarily bloating decision trees.
As a result we ended up with important and valuable assets distributed across the Tandeverse. We had ancient sales wisdom sealed in hidden Google Docs. We had powerful automation robots entombed within the subdirectories of that repo-of-a-few-projects ago. We had revealed truth on how to run projects better, faster, stronger contained within the gray matter of project managers. This is to say we had a lot of really valuable insights, works, discoveries and experiences that should have been immediately injected deep into the Flux Capacitor that powers the Tandem engine instead laying dormant; unused and forgotten.
Even still, there was a lot that went very right with Horoscope. It was able to power a good deal of our first iteration of Tandem. Looking back though, it's fairly easy to identify three pretty big things that we got wrong.
1. No no no! This one goes there! That one goes there! Right?
At it's essence Horoscope was basically a collection of markdown documentation. While not initially designed to be just that, a lack of clear guidelines about what kinds of things should live in it greatly hampered its utility and helped cause the aforementioned Diaspora of value.
If someone wanted to contribute important non-documentation assets back into the business it wasn't clear where they should do that.
2. I am a rock. I am an island
That said, and even if we had clearly defined what kinds of things should belong in Horoscope it existed on an island, detached from the day to day hustle of an agency. This meant there was no real way to disseminate materials contained within Horoscope so they were front in center in our most pressing work.
If someone wanted to add a useful cross-project script or guide back to the repo it wasn't clear how that asset could then be distributed to where it needed to go to be useful.
3. If you fail to plan; you plan to fail
The final major flaw of Horoscope was in how it figured out what to do next. An everything-goes submit-whatever paradigm was manageable when a decent amount of time was dedicated to sifting through things but it often produced disjointed tasks, impossibly hard to advance tickets like "redo our website" and a fairly subjective prioritization mechanism.
If someone had a great idea or wanted to surface a bug or proposal there was no standardized process to surface, evaluate, prioritize and translate them into reasonable chunks of work that could be done asynchronously over time by the entire team.