Back office
How do we move away from a software provider that keeps letting us down?
By Paul Meakin8 min read

Is it the provider, or the way the system was set up?
Start here, because the answer decides everything after it. Plenty of systems that "keep letting us down" were never set up for the way the business actually works. Fields nobody fills in. A process bolted on in the first week and never looked at again. One person who understood the settings, and then left.
Move that setup to a new provider and the problems come with it, plus an invoice for the move.
So before anything else, write down everything that has gone wrong in the last few months and sort it into three columns. The supplier's fault: outages, support that never answers, features withdrawn, price rises nobody warned you about. The setup's fault: wrong configuration, no training, a workaround that quietly became the process. And the ones nobody ever reported to the supplier at all. If most of the list lands in the second and third columns, a switch is an expensive way to rebuild the same mess somewhere new.
What does official guidance say about moving off an old system?
Some of the clearest guidance comes from government. GOV.UK's Service Manual guide to moving away from legacy systems opens by saying government uses a lot of older technology tied into costly contracts, and that yours probably is too. It is written for public services, but the advice travels to a small firm without much translation.
It says to agree a plan with the people the change affects, then set a date to turn the old system off. It is blunt about what happens without one: people avoid the final changeover, or renew contracts with the old supplier. It also suggests allowing a month or two for delays, but not so much slack that the deadline stops meaning anything.
The second warning is the easy one to miss. Do not pick a changeover date close to when the old contract or licence runs out. If the new system is late, you may have no choice but to renew the old one.
GOV.UK's separate guidance on managing legacy technology lists reasons organisations decide to move. Keeping the old technology costs more than replacing it. Supplier support is no longer available. The contract is due to expire. The risk of keeping it becomes too great. "We are fed up with them" is not on the list, so it is worth checking whether your frustration is really one of these reasons.
What are the options, apart from one big switch?
The same guidance names five approaches, and says which one, or which mix, depends on your size and how complicated your setup is.
| GOV.UK's approach | What it means in practice | When it tends to fit |
|---|---|---|
| Retain (do nothing) | Stay where you are, and fix the setup or the terms if that is the issue | Most of the problems were in the setup column |
| Retire (drop) | Stop using the system and the job it did | Nobody can say what the system is still for |
| Lift and shift | The same software, moved somewhere else | The product is fine, the hosting or support is not |
| Repurchase (shop and drop) | Buy a different product and move across | The product itself is the problem |
| Lift and shape | Move it and change it on the way | Some of it works and some of it needs redoing |
The first row deserves more respect than it gets. Staying put with a better setup and a renegotiated contract is a perfectly good outcome. Staying put because of what you have already spent is not. Sunk cost is not a strategy.
How do you switch without a big bang?
In stages. The legacy guidance says gradual retirement is generally cheaper and simpler than a full decommissioning project.
The Service Manual describes the strangler pattern. You put a wrapper in front of the old system, so everything else talks to the wrapper, and then replace the system behind it with little disruption to everyone else. For a small business the same idea is simpler. Move one team, one type of job or one set of customers first. Run it on the new system while everything else stays where it is, and compare the results against the old one before moving the next piece.
Test with real work. The guidance says to build something realistic rather than a simple model, so you can test under lifelike conditions. A demo full of tidy sample data proves very little. A week of your actual orders, with your actual mess, proves a lot.
And use the move to look again at how the work is done. The Service Manual says the changeover is the time to take a fresh look at your processes, rather than rebuilding the same system with new technology. Every workaround you carry across is one you will be paying for again on the other side.
What should you get hold of before you tell the provider?
Your data, and the paperwork that governs it.
GOV.UK's legacy guidance says the data and information in a system represents the true value of the system itself. It also lists the things that hold migrations up, including not knowing what data you have, missing documentation, and supplier contracts and exit arrangements. All three are easier to sort out while you are still on speaking terms with the supplier.
So test the export early. Find out what format it comes in, whether it includes attachments and history, who in your business is allowed to run it, and how long it takes. Do not assume it is instant, or that anyone can press the button. Find all of that out before the deadline, not during it.
Then read the contract for the notice period, the exit terms and what happens to your data when it ends. Make sure the admin accounts belong to the business, not to whoever happened to sign up first.
Where do people get this wrong?
Switching in anger. The worst week to sign a new contract is the week the old system fell over. That is when the demo looks best and the questions get skipped.
Tying the changeover to the contract end date, for exactly the reason the Service Manual gives. Leave room.
Treating software that has gone out of support as merely annoying. The NCSC's guidance on obsolete products says all software will eventually become out of date and that, ideally, once out of date it should not be used. It is plain that the only fully effective way to deal with the risk is to stop using the obsolete product. If your provider has stopped issuing security updates, that is a reason to move, not a grumble.
And rebuilding the same mess. If the first section's list was mostly setup problems, a new supplier will not fix them.
What should you ask before signing the next contract?
How long the product will be supported. The NCSC recommends considering the support lifetime of a product whenever you buy or license new software.
How you leave. What format your data comes back in, how long the export takes, whether it costs anything, and how much notice either side has to give.
What the supplier is planning. The legacy guidance tells organisations to consider how they manage supplier relationships and whether they know the supplier's future plans. A product being quietly wound down is easier to spot before you sign.
And how improvement is handled. GOV.UK notes that if you use a supplier rather than an in house team, you will have to negotiate continuous improvement into the contract. It will not appear there by itself.
We covered the checks on the product itself in what to check before buying another software tool.
When is this not the answer?
When the problem is the setup. Fixing configuration, training and process inside the system you already have is usually quicker and cheaper than a move. That is the work we do as business analysis and solution design.
When the question is legal. Notice periods, termination rights and disputes over service levels belong with your solicitor. The Service Manual itself says a team replacing legacy technology needs contractual, commercial and legal expertise alongside the technical knowledge.
When the system holds sensitive data and has gone out of support. That needs a security professional looking at it now, not a project plan for next year.
And when the honest answer is that the system is fine. If the list of failures is short and mostly minor, stay, and spend the money elsewhere.
Where do you start?
Write the three column list. Run a test export. Put the contract end date in the diary, and work back from it with room to spare.
If the list points to a move, or you cannot tell which column most of it belongs in, tell us what's stuck and we'll map the quickest fix. We will tell you straight whether that means a better setup, a staged move, or no move at all. Time to talk yet?
Common questions
How long should we run the old and new systems side by side?
Official guidance does not set a fixed period. Run each stage until the results on the new system match the old one for real work, then move the next piece across.
Can we leave before our contract ends?
That depends entirely on your contract's notice and termination terms. Read them before you plan dates, and ask a solicitor if the terms are unclear or you are in dispute with the supplier.
What if our provider has stopped supporting the software?
The NCSC says the only fully effective way to deal with the risk from an obsolete product is to stop using it. Treat it as a reason to plan the move now, and get security advice if the system holds sensitive data.
Do we need a developer to move our data?
Not always. Many systems have an export tool. Mapping the old fields onto the new system, checking nothing was lost and connecting it to your other tools is where help usually earns its fee.
Sources
Paul Meakin, Founder
Twenty years of fixing businesses from the inside, eighteen of them in recruitment from consultant to national operations, before building the automation, web applications and compliance systems Staxxd runs today.
More about Paul