User Adoption: Why Good Systems Fail After Go-Live
A system can pass every test and still fail, because people quietly go back to spreadsheets. The causes are usually specific, and most can be fixed in the first 90 days.

Six weeks after go-live, the new student information system at a network of private schools looks fine on paper. The data was migrated, the training sessions were held and the vendor has closed the project. Yet in three of the five schools, registrars are keeping attendance in Excel again “until the system settles down”, and fee follow-ups are back on paper lists.
This pattern is common, and it is rarely about the software being bad. Systems fail after go-live because the work around them was not designed for the people who use them every day.
Adoption is a design problem, not a training problem
When usage drops, the usual reaction is to schedule more training. Sometimes that helps. More often, people know how to use the system perfectly well and still choose not to, because using it makes their day harder. Training cannot fix a screen that needs eleven clicks for a task that used to take two, or a process that requires information the user does not have.
A more useful question than “do they know how?” is “what does the old way give them that the new way does not?”. The answer is usually specific and fixable.
Five reasons good systems get abandoned
1. The system was designed around the process on paper
Requirements were gathered from managers describing how work should happen. The real workday includes exceptions, interruptions and shortcuts that nobody mentioned. When the system has no place for those, users create one outside it. In the schools example, registrars handled late arrivals and half-day absences that the system’s attendance screen did not support, so they kept a separate list.
2. Nobody owns the system after the project ends
During implementation there is a project manager, a vendor team and weekly meetings. After go-live, all of that disappears. Small problems (a missing report, a confusing field, a permission error) wait weeks for someone to fix them. Each unresolved problem pushes a few more users back to their old tools.
3. The old tools were never switched off
If the old spreadsheet is still on the shared drive and the old system still accepts entries, people will use them under pressure. Parallel running has its place for a short, defined period. Without a firm end date, it becomes permanent, and the new system holds incomplete data that nobody trusts.
4. Managers do not use it either
Staff notice quickly whether their manager asks for reports from the new system or still requests the old Excel summary. If leadership keeps asking for the old format, they are telling everyone the new system is optional.
5. Data quality was poor from day one
If migrated customer records are duplicated, or opening stock balances are wrong, users lose trust in the first week. Once people believe the system is wrong, they verify everything manually, and the system becomes extra work rather than less.
What to do before go-live
The best adoption work happens before launch. Three things make the largest difference.
- Test with real users doing real tasks. User acceptance testing, the final check before launch, should be done by the people who will use the system daily, working through a normal day’s tasks, including the awkward cases. Their feedback on clicks and screens matters as much as whether calculations are correct.
- Train by role, not by module. A storekeeper does not need a tour of the whole inventory module. They need to receive goods, move stock and do a count, using their own items and locations in a training environment.
- Name a system owner and super users. One business owner who decides changes and priorities, and one or two “super users” in each team: staff with extra training whom colleagues can ask before raising a support request.
It also pays to look for small changes that make the new way easier than the old one. A default value that saves typing, a saved filter for the report a manager checks every morning, or a printed label straight from the receiving screen can do more for adoption than another training session. Ask each team for the three tasks they do most often, and make sure those three are fast. People forgive a lot in rarely used screens if the daily ones work well.
The first 90 days after launch
Treat the first three months as part of the project, with its own plan and budget. A simple sequence:
- Weeks 1 to 2: visible support. Super users and the implementation team are physically present or immediately reachable. Problems are logged and fixed daily, and fixes are announced so users see their feedback acted on.
- Weeks 3 to 4: switch off the alternatives. Set the old spreadsheets to read-only, close old system access for new entries and agree with managers that reports come only from the new system.
- Month 2: measure usage. Most systems can show who logs in and how many transactions are entered. Compare by branch or team. Low usage in one place usually points to one specific problem there.
- Month 3: review and adjust. Hold a short session with users from each team. Ask what is still harder than before and fix the top items. Then hand over to a normal support routine with a clear owner.
None of this is expensive compared with the system itself. Leaving it out is what turns a sound investment into a shelf of licences that nobody uses.
Systems people actually keep using
Adoption is earned in requirements, testing and the first months after launch, not at the go-live party. It depends on a system that fits the real workday, clear ownership after the project closes and leaders who use it themselves. The Launch stage of the KEMLAT Method treats this period as part of delivery, not an afterthought.
Our Enterprise Systems service covers requirements analysis, system design, data migration and post-launch support, with adoption planned from the first workshop and a named business owner in place before the vendor team leaves.



