The value of context
We thought we knew what we needed to build. Of course we needed and endlessly extensible search system to make sure we could build the perfect search algorithm. This would lead us to super charge our business and give our customers an unparalleled search experience which would help to drive revenue for the company. To do this we decided to onboard Solr and build views which would provide lightening fast results and a highly customizable search experience. The team put on their hard hats and got to the business of building an incredible system with lightning fast results with a fairly high maintenance cost knowing we would be able to play catchup later.
Fast forward 6 months and we are switching off our beautiful performant system and running off of a very simple home grown MySQL search engine with very few features integrating very little of what we learned along the way. How did we end up here? What could have gone differently to get to a better outcome? The missing magic sauce here was context. We thought we deeply understood the asked requirements but didn’t fully understand the context or the risk associated with what we invested in.
Context is the foundation to great decision making. If you don’t understand the environment around you and what is trying to be accomplished you won’t be able to communicate with your coworkers effectively or understand requirements when they are surfaced. Requirements and goals come out of context. When you have the context then you start to understand the “why” behind the KPIs, requirements, business decisions and behaviors within the organization.
To best understand context I like to ask questions. Get curious about why something is being asked for. How will you know if the team has succeeded. What are the biggest concerns and opportunities and why? Spend the time to develop measures from that context to help track how you are doing along the way. HBR just did a fantastic article about what happens if you focus too much on metrics which I recommend.
Once you understand the context more you can start to test ideas on what to do against this context. You’ll be able to see what isn’t relevant and what is crucial to success. Moving back to the search example I started with, we were missing a few pieces of context:
- An untested hypothesis that we needed more factors included in search
- Revenue and engagement would be strongly impacted by changes to search
- Success is measured by higher quality click throughs leading to conversations
With this added context we would have quickly tested some of these hypotheses to see if we needed to make the larger investment for Solr. This could have been 2 weeks of effort and could have saved us plenty of critical time.
As an aside, one of the great things that came out of building search on Solr is that we were able to build expertise so when we wanted to build a news feed experience we were able to do that fairly rapidly, but that’s another story…
