The product we are working on is a replacement for a successful existing product in our portfolio. It isn’t a good candidate for the strangler pattern since the main reason for the rewrite is that the UI platform it uses is deprecated and being removed from most major browsers soon. At the very least, the whole UI has to be swapped in one go - and that is the bulk of the app.

So what does delivery mean? We have a whole bunch of considerations:

• Do we have to be 100% feature compatible? I’m sure out of our entire user base there is at least one user for each wrinkle and featurette in the existing product.
• At what point do we start rolling out to users? MVP anyone? Not this time.
• Are users going to be able to switch forward and back between versions?
• Are we going to bake in new things that users have been clamoring for now, or later?

We have been wrestling with these concerns and were heading for a biggish bang release covering a significantly high coverage of the existing functionality, with a phased “beta” test period for some of our more trusted customers.

We are delivering to production infrastructure as part of our DevOps platform, so as to keep control over the process and not lead to the nightmare scenario of release day, which could be four months away, a big deal - since we’re fully automated and proven already and will keep it that way.

Even so, we are missing an ideal - that of genuine customer feedback.

Last week we got our chance. One of our smaller products is effectively a small subset of the overall product. A widget for third-parties to embed on their websites offering a live odds-comparison feed for upcoming races. We already have most of that working for the main product - so time to put our money where our mouths are and say “let us at it!”

This is a BIG deal for us. Going into production will help us focus on what matters. Hardening up the product and platform nice and early in our development process. Making sure we have our monitoring, logging, alerting, scalability, backups, disaster recovery, high availability, and so on, correct, and if not a chance to get them right before the main product launches.

This week, we did a minor pivot, and we now have a widget, almost ready to share with our partners.

Photo by Prateek Katyal from Pexels