
The CIAM Migration Diary · Part 6 of 6
The CIAM Migration Diary, Part 6. What changes once you have done it, and why the second migration is faster and safer than the first.
The question I am asked most since starting this diary is what would be different for another customer?
Simple, the next one will be even faster. Experience counts for a great deal, and we now have the tools to accelerate the work and to catch more things that can go wrong before they do.
On a first migration, most of the time does not go into building. It goes into finding things out. Which values consuming systems have fixed inside them. Which systems depend on each other. How the shared environments behave. What the operations team needs to see. Every one of those was a discovery this time. Next time, every one of them is method.
Much of it is already built. The compatibility work from Part 2, presenting the values consuming systems expect and producing a shorter token where a consuming product needs one, is now standard in our tooling. So are the runbooks, the dashboards, the migration measures, the chaos tests, the data conversion tooling and the deployment pipelines. The next customer is not building a platform. They are adopting one that has already been taken to production in a Tier 1 bank.
That changes what the programme is. A first migration is partly an engineering project. The second is an implementation.
During the programme we found hard coded values in consuming systems during regression testing, and dependencies between consuming systems when we ran complete journeys. Next time we will check for these aspects up front and observe what actually flows between the identity platform and everything around it before we build anything. We know the questions now, and asking them early costs almost nothing.
We will still build our own environment to behave like the customer's, as this has saved us more than any other single decision on this programme.
Testing order will change. Nothing moves until it is proven. Prove it on the incumbent first. One journey at a time, in a known order. Shared environments are treated as belonging to everyone else, because they do.
External dependencies are confirmed and held to account from the start: infrastructure capacity, other programmes' dates, third party changes. Part 4 showed this is where the real risk sits, and we now treat it that way from the first week.
Knowledge transfer starts with the build, as two programmes from day one, because the production team and the development team do different jobs. The customer's teams work on defects and configuration from the first week. Dress rehearsals against production data are in the plan from the start. And with the deployment pipelines in place, releasing a change is routine rather than a coordination exercise, which is where a great deal of time quietly disappears on a first programme.
Replacing a critical identity platform feels risky because of the unknowns. There will always be some unknown, but we now have the big ones in our playbook. That is why the next one will be faster, and safer.
This is not a projection made from a distance, either. Last week we ran a full dress rehearsal of the data migration, moving production data into the pre-production environment. The user migration completed in under an hour (it wasn't small but we had been optimising it over the preceding weeks). More importantly all of the data migrated successfully. Performance testing in the same environment showed the new platform running faster than the one it replaces (by a significant margin, in fact it is other infrastructure slowing us down).
There is still more to do. Failover and resilience testing come next, and the production waves start soon. The next instalment of this diary will come after go live (probably in the New Year), and it will tell you how it went. I am going into it confident, because most of the work that decides how it goes is already done.
If you want to see what a move like this could mean for your own organisation, our website explains how we do it, and our savings calculator puts your own numbers against it. And if AI is on your agenda, my other newsletter, Past the Easy Stuff, covers the business side of it.
If you want to change, or you are approaching a renewal and want to see what you could save, talk to us. We will tell you straight whether it makes sense for you, including when it does not.
The CIAM Migration Diary is Ajit Gupta's newsletter on LinkedIn. Subscribe on LinkedIn ↗

