Turnaround and recovery
When it has stopped working, and nobody can tell you why.
Sometimes it is a system: a crashed database, servers one failure from data loss, compromised accounts, a licensing deadline you are going to miss.
Just as often it is not the technology at all. Two teams keeping separate versions of the truth. A process that only exists in one person's head. Nobody able to say who signed off what. Costs nobody can account for. Those are turnarounds too, and they are the ones that quietly cost more.
What is actually broken is rarely just the software
When I am called in, the presenting problem is usually a symptom. The pattern underneath tends to be one of these three, and often all of them at once.
Systems
- Something has failed, or is one failure away from it
- Backups nobody has tested
- An application older than the people running it
- Tools that do not talk to each other, so people retype
- Spend that has grown by renewal rather than by decision
Process
- Work moving by email and memory rather than by system
- The same information keyed in three times, three ways
- Silos — site and office, finance and delivery, one country and another
- Nothing written down, so capability leaves when a person does
- A workaround from 2019 that everybody now treats as the process
Governance
- No clear owner for a decision, so it does not get made
- Nobody able to say who approved what, or when
- Suppliers managing you rather than the other way round
- Records that would not stand up if a client or auditor asked
- A board asking questions the business cannot answer
I look at all three, because fixing the server in a business whose real problem is that nobody owns the decision just buys you a quieter version of the same year.
What the first week looks like
-
Day one to three — find out what is actually true
Not what the documentation says, not what the last IT manager said. What the systems are doing right now, what is backed up and what is not, and what breaks next if nothing changes.
-
End of week one — the position, in writing
What has happened, what is recoverable, what is not, what it will cost and how long it takes. Costed and dated.
-
Then you decide
If you want somebody else to do the work from there, take the report and go. It is yours.
What an engagement covers
The common ones, in the order they usually happen:
- Stabilising what is failing now, before anything else
- Securing the environment — accounts, access, backup, and how the problem got in
- Recovering data and systems, in the order that gets the business trading again fastest
- Untangling how work actually moves between people, and fixing the handoffs that keep breaking
- Writing down what only exists in somebody's head, so the next person can run it
- Putting a decision owner and a record against things that currently have neither
- Working out what it should cost to run, and taking out what it should not
- Handing over something your own people can actually operate
That list is representative rather than exhaustive. In practice a turnaround can reach almost anything about how technology serves the business — who your suppliers are and what they are charging, whether two offices can work as one, how information reaches a client, which system should be retired, what to do about AI. If it affects whether the business runs properly, it is in scope. We agree the boundary in writing before starting, so you are never surprised by it.
How the money works
Scoped, priced and agreed in writing before any work starts. Fixed fee against a defined outcome — no hourly billing and no open-ended retainer.
If the work turns out to be different from what we scoped, we stop and re-scope it rather than run up a bill and hand it to you at the end.
Legacy systems are not a separate problem
More often than not, the thing that failed is old. An application nobody has the source for. A database written by somebody who left in 2009. A server that has been quietly holding the business together for a decade.
I have been working with these since before they were legacy — FileMaker, Lotus Notes, Access, on-premise line-of-business tools, bespoke systems built by a clever person who has long gone.
Most consultants will tell you to replace it. Sometimes that is right. More often you cannot afford the disruption, and the job is to make the old thing safe, supportable and able to talk to the new thing. I can usually tell you inside a week whether it should be repaired, wrapped, migrated or replaced — and what each of those actually costs.
Where I stop, and who picks it up
I do not personally run a helpdesk or a 24/7 desk. That is a deliberate choice about what I am worth paying for — not a gap you have to go and fill on your own.
Day-to-day support is available through my partners. If you need somebody to answer the phone when a laptop will not print, I have support firms I have worked alongside for years and will bring one in. You get the support, I stay involved at the level that needs a senior head, and you are not managing two suppliers who blame each other.
Deep specialist work is the same. Firewalls and penetration testing, heavy software development, construction-specific expertise — brought in from an associate network I have built over two decades, with me accountable for the result rather than handing you a phone number.
Licensed advice is not. Anything legal, tax, accounting or financial goes to somebody licensed to give it. I will tell you when you have reached that line.
So the honest summary: I am who you call when the problem is bigger than day-to-day support — and I can arrange the day-to-day support as well.
Proof
Twenty years with one client
Yearfull Group, a Hong Kong contracting group, have kept me on for two decades — for the decisions that carry risk, the upgrades that cannot go wrong, and the systems that did not exist until we built them. All of it around live projects that could not stop.
- 40%
IT running costs taken out, without the business losing a working hour.
- Zero
Downtime across server, network and firewall upgrades.
- HK$1.5m
Potential costs identified for avoidance on one tender by the AI checking system.
- Same day
Emergency recovery when a failure did happen; all systems restored inside seven days.
Something has stopped working?
Tell me what has happened. If it is not something I should be doing, I will say so and point you somewhere better.