For the past 10-15 years I’ve talked with fellow MSP owners about the “ideal IT support software”, and how great IT software could be… but over the years no one has dared to try and build it.
I think we’ve all known how challenging such software could be… And at the time of those conversations, I didn’t really know it would be me trying to build it! (Though I had my suspicions.)
Well here we are, and I think we’re getting very close…
Third beta release
We are targeting July 25th for the second bug-fix release of the Licorice beta provided everything is passed by our QA. This release aims to address the remaining integration issues. Other than a few edge cases for data synchronisation and the UI, everything is performing well.
Most of the bugs we’ve found when connecting to live ConnectWise instances in the cloud or on-prem have been unexpected things, like finding several integrations fighting over ConnectWise data fields… zero byte data stored in ConnectWise… super-long text strings that can’t sensibly be hyphenated… and certain types of scheduled events coming to ConnectWise from Outlook that we hadn’t anticipated.
These were all the types of “unknowns” we were expecting to find, and we can’t really know what they’ll be until we connect to real-world ConnectWise instances (even though we try our hardest to break Licorice during QA). So far as progress, many of the bugs are already addressed and in testing in the Licorice staging build. The remainder are scheduled to be addressed this current Sprint which ends on the 22nd.
I still expect there will be a few surprise to come, however the system is performing extremely well overall, all the data is intact and accurate within Licorice (even if some hasn’t sync’d), and the bugs are finite (even if sometimes it doesn’t feel that way!)
What we are prioritising for this period is what we’ve labelled “Data Consistency”. I.e. under no circumstances should Licorice be able to lose data; and even displaying data incorrectly could undermine confidence in the system—so this is our priority right now.
Our reasoning is that the most important thing of all, is to be able to trust Licorice with your live business data (even moreso than even the underlying software… i.e. ConnectWise). So issues relating to data consistency will be addressed first, along with bugs that block integration.
It is complex…
We were all correct in our assumptions… This has been difficult software to build. Why? Well, we have a whitepaper on why, and we’ll share some of this in time.
However the reason is largely, Licorice needs to operate far, far, faster than any underlying system it’s connected to. And to future-proof it, it needs to connect to and coordinate several systems (like Slack, Teams, Outlook, etc.)
This means first of all our servers need to respond to anything in milliseconds, and the UI needs to give meaningful feedback in under 100ms. This also means often we have to make decisions based on incomplete data from ConnectWise (and other systems)—and yet we also have zero tolerance for errors in that data: It HAS to be 100% accurate.
This is a complex area of computer science, but, the good news is, it’s done. Licorice already has these systems underpinning it, and they’re working very reliably. The details we’re dealing with above are in the final wirings to and from ConnectWise.
Licorice also works in such a way that you can jump around between jobs at speed and never lose information—even if you were part way through getting a thought down. This is just a reality of the job… IT support moves very quickly sometimes, yet requires 100% accuracy, and 0% error tolerance. If you miss one piece of software out of 45 on someone’s computer, it will be noticed, and need to be corrected.
This is why we need software that can move as fast as we think while we’re doing this work. Not software that slows us down. We also need software we can depend on to keep 100% accurate data, even if the software it connects to goes offline. That’s just part of the job… So that’s what we’ve built.
We have worked with the first 10 beta partners and will soon have the blocking bugs addressed, and have also captured a lot of crucial feedback from those sessions.
Once Licorice is working well for this first group, we will open up the next group to 20 ITSPs/MSPs to join. There will be a few more bugs, which we’ll again resolve, then I expect we’ll be ready to open up to the 40 ITSPs/MSPs.
At some point in the coming months, we will stop seeing bugs. I call this method “gating”. I would much rather do things this way, than tell everyone we’re ready and ultimately just end up frustrating a lot of people. We instead take an engineering approach being: First a small group, then a bit bigger, then bigger, as we gain confidence and tune the software.
I also am leaning towards taking a leaf from Fedora’s release policy, which is that if a release date is missed due to a QA problem (which invariably happens in software) then we will automatically attempt to release 1 week from that date. So for example if we don’t pass QA for July25th, we will automatically re-focus on August 1st.
Though we will always try to hit the first release date, this system gives additional predictability and clarity around release in situations that would otherwise be unclear.
Want to talk?
I’d like to thank our beta partners for their fantastic input, and trust in us, and as always our phenomenal team who work tirelessly to make the world a little bit better for IT engineers everywhere.
As always I’m here on our Discord if you have any questions and suggestions, and Discord is where announcements (like this) are posted first.
This is the software we always wanted at our MSP and with your help, we will make sure it’s the software we’ve all always talked about.