Agree @edgescape, I've reached out to management to try and get more details on cost/time to achieve those efficiencies. Will report back if I hear anything of note.

In my experience, the internal hosting costs often are exacerbated by poor management by the development, test and support teams. For single tenant cloud architecture for each customer there will be environments (virtual machines, storage etc) for production (24x7), pre-production (to test the installation process), user acceptance test (2 weeks per release), system test (several weeks per release), training, support (to replicate support issues), and development environments. So you can imagine the Amazon or Microsoft costs for all of this every time you win a new customer! As a rule of thumb to scale effectively, the hosting costs should be 1/7 to 1/10 of the related SaaS fees.

There are likely some environments running overnight when the teams don't need them and it takes a bit of fiddling around to set up a power on/off schedule for possibly dozens of different virtual machines. The teams like the convenience of the the environments on all the time because it saves them having to think about when they might need access and saves waiting for a start up period if an environment is suddenly needed for something. However, the teams are often not responsible for the costs of these environments so there is effectively an incentive for them to waste hosting costs. Unless there is a cloud operations manager tasked this monitoring all of this, and planning an automated process for cycling on/off at the appropriate times for each team then it's very hard to address.

Also, there may be the option to "multi-tenant" the environments so that they can be shared by more than one customer. This requires a much more advanced architecture to ensure the customer data doesn't get mixed up causing a data breach, and to allow the external interfaces to interact independently to operate in parallel.

This is a long way of saying, it may take a while to bring down the hosting costs. Depending on the issue, there could be some quick wins in 3 months, then other things may take more than a year.

This is a good level of insight @actionman. I’m curious, is the approach you lay out one that would be obvious to Prophecy management already, or would they appreciate an experienced shareholder pointing it out?

This is a brilliant post @actionman and highlights the benefit of the Strawman community. It reads as if the multi-tenancy solution is being looked at longer term, but to your point it will take time:

I suppose from an investment point of view it is more about the business scaling over the revenue growth. Any measure to bring down the absolute cost of hosting is nice, but most importantly whether it's a tighter focus on managing the current setup or moving to a new one over time, the key is ensuring hosting costs increase at that 10-15% rate relative to revenue as you point out. Not the 60%+ rate it did this year!

Thanks @lastever I'm sure they will be aware of the issue but it's a matter of competing priorities for management to focus on and it's a difficult problem to solve. In some way focusing reducing hosting costs is a distraction from scaling which will prioritise sales (as seen in their commission payments jumping up), deploying to new customers, and building the product suite out. So hosting costs may come in 4th priority. Also, there are dependancies. The product and development environment may have some uplift required to solve some of the hosting issues. Plus they need to have staff available to take on the cloud operations role which is difficult to recruit. None of this is easy. They will need to solve it eventually because it is free money which will continue to grow, and will hold them back.

@Wini Yes, multi-tenancy would be ideal. For example, Xero does this. I think Xero took 6 years to rearchitect and migrate to AWS and achieve this (plus other improvements no doubt).

Some software is very difficult to achieve multi tenancy, I.e. where it has a high degree of bespoke configuration and integrations. Again, Xero doesn't have this problem because the software is following a standard business process governed by rules set by the ATO, accounting standards and so on.

So the question for eMite and Snare is, can they actually share infrastructure without rearchitecting the products? I don't know enough about their solutions, but I suspect their customers will have some bespoke requirements (rules, business process, and integrations) that are mandatory requiring customisation and cannot be omitted or worked around with a more generic solution. Integrations may be tightly coupled two-way, realtime so it may not be easy to abstract that away into a shared middleware architecture. If you ask management they will be upbeat and say of course they can do it, but they may not truely know, or want to know, that level of detail. It may be possible to deduce that from public information though. But, I guess we will see in their next reporting season.

They may decide to only multi-tenant some newer or simpler customers, and have legacy customers on single tenant.

If you are doing modelling, AWS costs approx A$50k pa for a typical single-tenant instance using AWS RDS, app servers, storage etc, which is 1 production environment, then 3 other scheduled power cycled environments for non prod (dev, test, UAT, training etc) and that would suffice for one product upgrade cycle per year. Maybe something like A$70k for high availability. If they are using Oracle database or Microsoft SQLServer they will be paying a double or triple that and trying to move to AWS RDS database ASAP which sounds like their medium term plan before considering multi tenant. I.e. the $700k MS licensing.

@Wini @actionman

I was thinking about the same thing when comparing PRO to what GTK had gone through when the new CEO came in to save the company after a period of underperformance in the platform before the share price had bagged multiple times in short period of time

To me this looks like a "lightbulb" moment. But before I get excited at looking at a potential Gentrack and opening up my wallet to Prophecy like lots of others here have done, there are a few key differences.

1. Gentrack hired CEO Gary Miles with a deep development background at Amdocs. Prophecy CEO doesn't have this deep experience. He seems like more of a marketing guy.

2. Gary Miles used his networking in the dev world and able to bring his "A-team" in to quickly fix the problems at Gentrack. Hell, Miles even brought in a guy from Xero! The question remains whether he can pull in top talent in the industry

I highlighted this as a risk before that top talent is hard to get and I was not sure if Prophecy had the A-team to scale their products in what looks like a fragmented industry.

Multi-tenant solution (ie: docker containers etc...) is good but it is hard to implement until you understand the concepts. I find it hard too and I'm still learning about it.

As mentioned before, I think this will take time given the need to hire top talent, but willing to watch this develop and sit this one out.