Implementation · Adoption

Live on Odoo, but nobody uses it: fixing adoption after go-live

Go-live is signed off, but the spreadsheets are back and managers have stopped reading the reports. Here is how to tell an adoption problem from a configuration problem, how to measure use with data Odoo already keeps, and the order in which to fix it.

By Octagotech Team9 min read

Key takeaways

  • Run one task end to end first. If it works when followed, the problem is adoption, not configuration.
  • Measure use with what Odoo records: latest sign-in, records per user, overdue activities.
  • Restore trust in one disputed number before you retrain anyone.
  • Train by role on daily tasks, with a super-user in each team.
  • Retire the parallel spreadsheets on a set date, then review every month.

Go-live happened. Invoices go out of Odoo, the project has been signed off, and the old system is switched off. Yet the warehouse keeps its own stock sheet, sales managers still ask for the weekly numbers in Excel, and half the team signs in only when someone chases them. On our Odoo rescue page this is the sign we call "live, but nobody trusts it". It is the quietest of the seven warning signs, because nothing is broken enough to raise a ticket.

This guide is for teams in that position. It covers how to tell an adoption problem from a configuration problem, the most common causes, how to measure use with data Odoo already records, and the recovery sequence we follow.

Adoption problem or configuration problem?

The two look alike: people avoid the system and the numbers look wrong. The difference is whether the system does the job when someone uses it properly. A configuration problem means the workflow itself is wrong: a route that never creates the delivery, a tax that posts to the wrong account, an approval step that blocks orders it should let through. An adoption problem means the workflow works, but people skip it, use it late, or use it halfway and finish the job somewhere else.

The quickest test is to sit next to someone and have them run a normal task from start to finish, with the process owner watching. If the task completes correctly when the steps are followed, and the result matches what the business expects, the configuration is sound and the problem is adoption. If it fails, or produces a number nobody can explain, fix the configuration first. Training people harder on a broken flow only teaches them that the system cannot be trusted. Often it is both, so fix what gives wrong results before working on how people use it.

A second test: when a manager wants stock on hand, open orders or this month's margin, where do they look? If the answer is a spreadsheet or one person's memory, the system has not been adopted, whatever the login count says.

Why people stop using Odoo after go-live

These are the patterns we meet when we are called in to restructure projects other vendors started. Often more than one is at work.

Training covered features, not tasks. Users toured the Sales app, then went back to a job that starts with a phone call and ends with a delivery. Nobody walked them through the screens their day passes through, in order, with their own products and customers. People who cannot finish a task alone go back to what they know.

Screens carry fields nobody needs. Forms were built to hold every field anyone mentioned in a workshop. A buyer who needs a handful of fields faces a screen full of them, several required, and has to guess which ones matter. Every unnecessary field slows entry down and invites wrong data.

A report disagrees with the old spreadsheet. This one does the most damage. In the first weeks a manager compares Odoo's sales or stock figure with the spreadsheet they trust, the two disagree, and nobody explains why. The difference may be legitimate, such as a cut-off date or a valuation rule, or it may be an error. Either way, once a manager has seen a number they cannot reconcile, they go back to the spreadsheet, and their team follows.

Approvals route to nobody. An approval rule names someone who has left, points at a group nobody was added to, or waits on a manager who never learned where approvals appear. Documents sit, the requester gives up and does it outside the system, and the workaround becomes the process.

Views are slow because of custom code. A list that is slow to open, opened dozens of times a day, is enough to push people back to spreadsheets. The usual cause is custom code: a computed field that runs a search for every row, or an automated action that fires on every save. Odoo's built-in profiler, switched on from the developer mode tools, shows where the time goes.

Nobody inside owns the system. The partner handed over at go-live and no one internal was named to own Odoo. Questions go unanswered and each department settles on its own workaround. Without one internal owner who can make decisions, none of the fixes below will stick.

Measure adoption with what Odoo already records

Before changing anything, find out who uses the system and for what. You do not need an analytics tool: Odoo keeps enough on every record, and a baseline taken now lets you show progress later.

Start with sign-ins. The Users list in Settings shows each user's most recent sign-in, in a column called Latest authentication up to Odoo 18. In Odoo 19 it is called Latest Login and is an optional column you switch on from the list's column toggle. It records when someone last signed in, not what they did afterwards, so treat it as a first filter: licensed users who have not signed in for weeks are either not using Odoo or sharing someone else's account.

Then count the work. Every Odoo record stores who created it, and any list can be grouped by that field: open the list, choose Add Custom Group in the search options and pick Created by. Do this for the documents each team should be producing, such as quotations, purchase orders, receipts, manufacturing orders or vendor bills, over the last month. A salesperson with no quotations in Odoo is still quoting somewhere. If one person creates every record for a team, you have found someone keying in work for colleagues who never learned the system.

Look at overdue activities next. Odoo marks overdue activities in red in the chatter and activity views. In Odoo 18 and 19, developer mode adds Settings ‣ Technical ‣ Activities ‣ Activity Overview, which has an Overdue filter. Group it by Assigned To and you can see who has stopped working their queue. A long tail of overdue activities on one user usually means that person does their follow-up somewhere else, or has left and nobody reassigned the work.

Finally, read the chatter on a sample of records. Fields that Odoo tracks, such as the status of an order, log who changed them and when. If orders jump from draft to done at the end of the month, the work happened elsewhere and was keyed in afterwards. Then collect the spreadsheets people still keep, and note what each one does that Odoo does not.

SignalWhere to find itWhat it tells you
Latest sign-in per userSettings, Users list: Latest authentication (Latest Login in 19)Who has stopped opening Odoo
Records created per userAny list, grouped by Created byWho works in Odoo, and who has someone else do it
Overdue activities per userActivity Overview in developer mode (18 and 19), grouped by Assigned ToWhose follow-up happens somewhere else
Status changes in the chatterA sample of orders, bills and transfersWhether work is entered as it happens or keyed in later
Spreadsheets still in useAsk each team for themWhich answers people do not trust Odoo to give

A six-step recovery sequence

Order matters more than any single fix. Retrain people before the numbers are right and you teach them a system that still gives wrong answers. This is the sequence we follow.

  1. Restore trust in one number first. Pick the figure the leadership team argues about most, often stock on hand, open receivables or the month's sales. Reconcile Odoo's figure with the spreadsheet line by line, fix what is wrong in Odoo, and write down the differences that are legitimate. Then walk the people who stopped trusting the report through the reconciliation. Comparing key reports with the legacy output is on our pre-go-live checklist for exactly this reason.
  2. Simplify the screens people use every day. For each role, list the forms and lists they open daily and remove what they do not need. On Enterprise, Studio can make a field invisible or show it only to certain user groups. On Community, a small inherited view does the same. Mark secondary list columns as optional so users can hide them, and keep required fields to the ones the process truly needs.
  3. Retrain by role, with a super-user in each team. Train each role on its own daily tasks, in order, with its own data. Name a super-user in every team: someone who uses Odoo daily, gets deeper training, answers first-line questions and takes problems to the internal owner. A question answered at the next desk does not turn into a workaround.
  4. Write procedures for repeated tasks. One page per task, with steps and screenshots from your own database: receiving a delivery, raising a credit note, closing a manufacturing order. Keep them where people work, in the Knowledge app on Enterprise or a shared folder, and have the super-user update them when the process changes.
  5. Retire the parallel spreadsheets on a date. Announce the date, and make sure every question a spreadsheet answered can be answered in Odoo by then, through a saved filter, a report or a dashboard. On the day, the spreadsheet becomes read-only. If one survives because Odoo cannot answer its question, treat that as a configuration gap to close, not a habit to scold.
  6. Review monthly. Repeat the baseline measurements every month with the internal owner: sign-ins, records created per user, overdue activities, and which spreadsheets are still alive. Keep it going after the numbers improve, because new starters and process changes bring old habits back.

One reconciled number beats a training day. Managers go back to spreadsheets when a figure disagrees and nobody can say why. Explain the difference in writing, fix what is wrong, and they have a reason to come back.

Where this sits in an Odoo rescue

On our rescue engagements this is not a separate project. The Diagnose phase, which comes before any quote, audits how people use the system alongside the custom code, the data and the integrations, and the measurements above are part of it. The root cause is named in writing, because adoption that failed through slow custom code needs a different fix from adoption that failed through missed training. Our note on the rescue diagnostic audit describes what it covers.

Stabilize comes next and fixes what stops daily work first: the approval nobody can act on, the view that takes too long to open, the number that does not reconcile. Restoring trust in one number starts here when the wrong figure stops daily work; the wider data reconciliation belongs to Repair. The Repair phase runs on a dated plan and covers the rest of the sequence: screens simplified, custom code repaired or replaced, data cleaned and reconciled, users trained by role with a super-user in each team, procedures written, and spreadsheets retired on their date.

Then comes Stay: hypercare once the repaired system is live, then a support tier with our Rajkot desk, which is briefed on your system before it takes the first ticket. The monthly review fits there, and Odoo support with 24/7 on-call for critical issues is available on the Enterprise tier. If the database is also on a version outside Odoo's standard support, the diagnosis covers that too, and what drives the cost of an Odoo upgrade explains how we size it.

What we need to start is short: admin access to the database and hosting, the source code of every custom module, whatever scope documents exist, and one internal owner who can make decisions. The recovery is quoted after the diagnosis, at a fixed price wherever the scope can be pinned down.

Frequently asked questions

How do we tell an Odoo adoption problem from a configuration problem?

Have someone run a normal task from start to finish while the process owner watches. If it completes correctly and the result matches what the business expects, the configuration is sound and the problem is adoption. If it fails, or gives a number nobody can explain, fix the configuration first.

Can we measure Odoo user adoption without extra tools?

Yes. The Users list shows each user's most recent sign-in, any list can be grouped by Created by to count records per user, and overdue activities show whose follow-up happens somewhere else. Take a baseline before you change anything and repeat it every month.

Should we retrain everyone after a difficult go-live?

Not with the same course. Retrain by role, on the tasks each person does every day, with your own data, and name a super-user in each team. Do it after the disputed numbers reconcile, or you are training people on a system they still have reason to distrust.

How long does it take to fix adoption after go-live?

It depends on what the diagnosis finds. Wrong numbers that stop daily work and approvals that reach nobody are fixed first, in the Stabilize phase, and the wider data reconciliation follows in Repair. Simpler screens, retraining and retiring the spreadsheets follow on a dated plan that is agreed and quoted before the work starts.

Octagotech Team

Certified Odoo Ready partner · San Francisco, New York, Ahmedabad and Rajkot

Our New York project managers and Ahmedabad engineers diagnose and repair Odoo systems that other partners started, and our Rajkot desk supports them afterwards. About Octagotech, Odoo Ready partner.

Live, but not trusted?

30 minutes: tell us which number your managers stopped believing, and we will tell you where we would start.

Book a discovery call

Find out why your team went back to spreadsheets

The first call takes 30 minutes.

Or email sales@octagotech.com