I’m writing this post to share what we’ve been up to in the first months of 2023. Taking a look inside Compiuta might be interesting to someone and, at the very least, this will be another piece of our logbook.

Main events

This quarter we took part in A&T, an automation and testing fair in Turin, Italy. We decided to join last minute (less than a month before it took place) and committed all our resources to:

• update our existing Connhex flyers, to better highlight its strengths. More on this in the following section
• create a roll-up to explain what Connhex actually is (no, it’s not an IoT platform!)
• create a video about Connhex: we kept looping it for the entire fair
• update Connhex website’s homepage
• revamp our existing demos
• ditch Google Analytics in favor of Plausible Analytics
• get logistics right

As you can see, we went all in with Connhex. Explaining that Compiuta is the company whereas Connhex is the product was a bit cumbersome at times, but we did our best.

Doing all of this took us approximately a month. The website revamp was particularly time-consuming since we:

• had to start from scratch to propose Connhex under a new light
• had to come up with a new layout
• wanted to be proud of the final result. The previous version of Connhex’s website stayed online for way more than it was intended to

There was a risk involved in undertaking this a few weeks before the event: we would have bitten more than we could chew. We decided to go ahead nonetheless and managed to get the homepage about 95% done: instead of taking down every deprecated page, we placed a banner on top of those apologizing for being late.

Product

If you have been following us along for the ride, you already know about our decision to go all in on Connhex.

Connhex has evolved a lot during 2022, so it was time to redefine what its capabilities are. After a lot of reframing (maybe a topic for another post) we settled around:

• focusing on hardware manufacturers
• separating the offering in three: Edge, Cloud, AI
• rethinking our pricing

We have been busy benchmarking how the IAM service scales: it was necessary to go deep down the rabbit hole to optimize how policies are evaluated, but now we’re happy with it. Furthermore, we have been testing additional notification methods (we had only relied on email in production use cases, up until this point) and improved Connhex Auth by adding support for verification codes.

We also dedicated some time to squash a few bugs in Connhex’s rule engine and removed the requirement to specify an event duration.

Alongside with developing Connhex, we also worked on two custom projects (based on Connhex itself, obviously) and performed a round of maintenance on the instances we host.

Likes and dislikes

When I started writing this post (1 hour and 53 minutes ago) I assumed the “what went wrong” section would be much longer than the “what went right” one. I’m not so sure about that anymore: we were able to move forward on multiple directions in spite of underestimating how long the preparations for A&T would take.

We were unable to improve Connhex’s online presence. We have so much stuff built, tested and deployed in production that we don’t even mention on our website and docs. If I were to visit Connhex’s website without having access to all of our git repos, it would feel like a fake product. When we show Connhex’s features in person, the typical reaction is awe. Now, I would really like to think we have built something revolutionary: the most probable thing, however, is that our online presence set the bar way too low.

No worries though - just something to work on next quarter.

Next quarter

Our priority for Q2/2023 is to improve Connhex’s online presence. We plan to tackle this by:

• polishing our internal documentation and moving it online
• describing all the services that make up Connhex and what you can use each one of those for. Fortunately, recreating the homepage gives us a good foundation to build upon
• adding use cases and sharing how customers are using Connhex

The last point is proving tricky. We would like to do this properly, by having customers actively participating instead of simply putting words in their mouths: we’ll see if we can pull it off. Some of them shared their fear in having competitors discovering their use of Connhex and catching up quickly with them, so we’ll see how we can handle this.

We have a Connhex-based project that is almost finished: we will be focusing a lot of our energies on it, making sure everything runs smoothly once moved from staging to production.

On the commercial side of things, we’ll be starting an effort to scale Connhex’s distribution through software houses. We already took an unsuccessful stub at this, but we think we should have been explaining what advantages using Connhex would bring better.

And yes, we’ll be exploring a GPT-based integration inside Connhex, because one should always make time for fun.