The 7 Most Common Mistakes in ERP Implementation

Short answer: The seven most common ERP implementation mistakes are: no clear goal, starting without process analysis, not involving employees, over-customization, neglecting training, ignoring data quality and choosing an inexperienced partner. The cause is rarely the technology and almost always the approach: these mistakes stretch timelines, inflate budgets and leave staff working around the system instead of in it. To avoid them, start from measurable goals, split the project into clear stages and work with an experienced partner who knows local conditions.

Companies usually start an ERP project with high hopes: order in the books, accurate stock, real numbers for management every day. A few months later the deadline has slipped, the budget has grown and employees still keep Excel sheets next to the new system. A share of ERP projects fails to deliver the expected results, and the culprit is usually not the technology but the wrong approach.

Below are the seven most common ERP implementation mistakes, each with warning signs, cost and prevention, plus a checklist. The implementation phases themselves are covered in our Odoo ERP implementation guide; here we focus on the decisions that break a project.

What counts as an ERP implementation mistake, and why is it expensive?

An ERP implementation mistake is a wrong decision, made at the start or during the project, that damages the timeline, the budget or how much people actually use the system. Such mistakes are made not in the software but in how the project is run.

The later a mistake is found, the more it costs: an hour of discussion during analysis, rework after configuration, wrong documents and lost trust after go-live. That is why the warning signs below matter most.

The seven most common mistakes:

1. No clear goal: starting the project without answering "why do we need an ERP".
2. Starting without process analysis: configuring the system without understanding the existing problems.
3. Not involving employees: when the team's input is ignored, resistance to the system only grows.
4. Over-customization: spending time and budget on features nobody needs.
5. Neglecting training: employees never learn to use the system properly.
6. Ignoring data quality: migrating old, inaccurate data into the new system.
7. Choosing an inexperienced partner: working with a company that doesn't know the local market or offers no support.

Mistake 1: Starting an ERP project without a clear goal

A clear goal is a result expressed in numbers and dates: what should change in the business once the ERP is live. "Automate everything" is not a goal.

Warning signs:

• the specification lists modules but no expected results;
• nobody can say when the project will count as a success;
• new requirements keep arriving.

What it costs: without a goal, scope grows endlessly because there is no basis for saying "we don't need that". Time and budget go up, the effect can't be measured after go-live, and management loses faith in the project.

How to prevent it:

1. Write down 3–5 measurable goals, for example "shorten the month-end close" or "see sales and receivables in one daily report".
2. Record today's value for each goal so you have something to compare against.
3. Appoint a project sponsor (usually the director or owner) who has the final word on conflicting requirements.
4. Put the processes that control money and goods into phase one; the rest can wait.

If you are still deciding whether you need an ERP and which one, start with 10 key factors when choosing an ERP.

Mistake 2: Configuring the system without process analysis

Process analysis means documenting how work is done today (AS-IS), where time and money are lost, and how it should work in the new system (TO-BE).

Warning signs:

• configuration starts from demo defaults because "we'll adjust it later";
• employees describe the same process differently (for example, customer returns);
• exceptions such as discounts or sales on credit are not documented anywhere.

What it costs: a system built on undocumented processes simply automates the current mess. Configuration is redone, testing drags on, and whatever the system doesn't cover moves back into Excel, so the reports never match.

How to prevent it:

1. Map the core flows end to end: order → payment → shipment → invoice; purchase → receipt → payment.
2. For each step, note who does it, with which document and in which software.
3. Mark the bottlenecks: re-entered data, waiting, frequent errors.
4. Get department heads to approve the target process, and keep it as close to the system's standard logic as possible.

For where to begin, see Business process automation: where to start.

Mistake 3: Not involving employees

Involving employees means bringing the people who will use the system every day into analysis, testing and decisions. Management picks the ERP, but warehouse staff, cashiers, sales managers and accountants work in it.

Warning signs:

• staff first hear about the system at training;
• no key users have been named, or they have no time for the project;
• documents are entered late or not at all, while notebooks and Telegram chats carry on in parallel.

What it costs: the system is paid for but only half used, management distrusts the incomplete figures, and in the worst case the company returns to its old tools.

How to prevent it:

1. Name a key user in each department and free up part of their workload for the project.
2. Explain at kick-off which manual work disappears and whose job gets easier.
3. Bring key users into analysis and testing: people resist a process less when they helped approve it.
4. Set the rule at management level: if a document isn't in the system, it doesn't exist.

Mistake 4: Over-customization

Over-customization is using custom code to bend the system to old habits where standard features would do the job. Customization itself is fine; unjustified volume is the problem.

Warning signs:

• dozens of "make it like the old system" requirements, such as copying 1C print forms exactly;
• "our business is unique" is said before anyone has looked at the standard process;
• the plan gives more time to programming than to configuration and training.

What it costs: you pay for each customization three times: to build, test and maintain it. Odoo releases a new major version every year, and custom modules must be adapted at each upgrade. More customization means more expensive upgrades and more dependence on a single developer.

How to prevent it:

1. Adopt a "standard first" rule: show the standard feature before discussing custom work.
2. Run a fit-gap analysis: sort requirements into "standard", "configuration" and "development".
3. Justify each customization in money or time saved.
4. Move nice-to-haves to a later phase; after go-live, many turn out to be unnecessary.

On adapting a ready-made system versus building your own, see Building an ERP from scratch or using a ready-made system.

Mistake 5: Neglecting training

ERP training means preparing each employee to work in the system independently, for their role, on their own data, before go-live. One general presentation is not training.

Warning signs:

• one 2–3 hour session is planned for all departments;
• there are no instructions, or they are in a language staff don't work in;
• after go-live the partner is flooded with basic questions.

What it costs: untrained staff create wrong stock moves, prices and invoices. Each error takes the accountant's time, erodes trust in the reports and pushes support costs up.

How to prevent it:

1. Plan training by role: sales, warehouse, cashiers, accounting and management need different things.
2. Train on a test database with the company's own products and customers.
3. Prepare short instructions and videos in the languages your team uses (in Uzbekistan, usually Uzbek and Russian).
4. Use "train the trainer" so key users can onboard new hires.
5. Before go-live, check that each person can do their daily tasks without help.

Mistake 6: Ignoring data quality

Data migration means moving customers, products, stock balances and open debts into the new system after cleaning and checking them. Migrating old, inaccurate data makes the new system unreliable from day one.

Warning signs:

• one product exists under several names (Latin and Cyrillic, different abbreviations);
• the customer list has duplicates, and some counterparties have a missing or wrong taxpayer ID (STIR);
• MXIK (IKPU) product codes are incomplete;
• opening stock is taken from the old program without a physical count.

What it costs: with wrong balances, the system lets you sell stock you don't have or hides stock you do. An e-invoice (for example, via Didox) with a wrong STIR or MXIK code can be rejected, and duplicates distort reports. Soon nobody trusts the numbers.

How to prevent it:

1. Clean data first: merge duplicates, agree a naming standard, archive inactive records.
2. Migrate only what you need: usually active master data, opening balances and open debts.
3. Count stock as of the go-live date and reconcile opening balances with the accountant.
4. Run at least one test migration and have key users check it.
5. Fill in tax-critical fields in advance: STIR, MXIK code, VAT rate.

Moving from 1C? See Migrating from 1C to Odoo.

Mistake 7: Choosing an inexperienced partner

An implementation partner is the company responsible for configuration, customization, data migration, training and post-go-live support. A partner who doesn't know the local market or offers no support makes every mistake above more likely.

Warning signs:

• the partner was chosen on price alone;
• they can't show clients in your industry;
• they have no clear answers on Uzbek requirements (12% VAT, e-invoicing, Asl Belgisi marking);
• post-go-live support isn't in the contract.

What it costs: projects stall halfway or are handed over "working, but wrong". The company then hires another partner to redo part of the work, paying twice for the same job.

How to prevent it:

1. Check the partner's status in the partner directory on odoo.com.
2. Talk to their clients in a similar industry and ask to see a live system.
3. Ask about their methodology: phases, deliverables and who is needed from your side.
4. Check for ready local solutions: Didox, Soliq, Asl Belgisi, migration from 1C.
5. Put support terms in writing: response times, channel, price.

The full list of questions is in How to choose an Odoo partner in Uzbekistan.

How are these mistakes connected?

The seven mistakes form a chain: without a goal, analysis stays shallow; shallow analysis leads to over-customization; and customization eats the time meant for training and data. Fixing one early often prevents the others.

Signs a project is heading the wrong way:

• phase results aren't signed off;
• the requirements list grows every week;
• the go-live date has moved more than once;
• migration and training sit at the very end of the plan.

If two or more apply, pause and revisit goals and scope. On how project cost is built up, see How much does an Odoo implementation cost.

ERP implementation checklist: what should you check?

A checklist is a short list that tells you, at the end of each phase, whether you are ready to move on.

Before the project starts:

• 3–5 measurable goals and their current values are recorded;
• a sponsor and an internal project manager are appointed;
• core processes are mapped and approved by department heads;
• each department has a key user with time for the project;
• the partner's experience, methodology and support terms are checked.

During configuration and testing:

• every customization has passed fit-gap analysis;
• data is cleaned and a test migration is done;
• key users have tested real scenarios;
• STIR, MXIK and VAT fields are filled in.

Before go-live:

• role-based training is done and instructions are ready;
• stock is counted, and opening balances and debts are reconciled;
• a shutdown date for the old system is set;
• the support process after go-live is agreed.

How do Odoo and Deep Vision help avoid these mistakes?

Odoo is modular, so you can start small and extend step by step; Deep Vision runs the project in clear stages and accounts for local requirements. No system prevents mistakes by itself, but good structure makes them far less likely.

• Phased go-live. We start where money and goods move. Standard Sales and Inventory often go live in 6–8 weeks; broader projects take 2–6 months.
• Standard first. We show Odoo's standard features first and propose custom work only where it pays off.
• Ready local modules. We have ready Odoo modules for e-invoicing via Didox and for Asl Belgisi marking.
• Training and support. We train staff by role and keep supporting the system after go-live.
• Experience. Deep Vision is an official Odoo Gold Partner with 7 certified specialists and 26 client projects in the Odoo partner directory. Clients include Optika.uz, Timsoll and RockBeton.

Deep Vision works to a proven methodology. See what the service includes on the Odoo implementation page, or book a free consultation where we assess your project's risks together: leave a request.

Frequently asked questions

Why do ERP projects fail?

Usually because of the approach, not the technology: no clear goal, no process analysis and no employee involvement.

Which ERP implementation mistake is the most expensive?

Usually over-customization and an inexperienced partner. The first adds cost at every upgrade; the second forces you to redo work.

Should we avoid customization completely?

No. It makes sense where it gives a real advantage or meets local requirements, as long as each piece is justified and documented.

What if employees resist the new system?

Find the reason, usually confusion or fear of extra work. Involve key users, add hands-on training and make "only documents in the system count" a rule.

Do we need to migrate all our old data?

Usually not. Active customers and products, opening balances and open debts are enough; older history can stay in an archive.

Can a troubled project be rescued?

Often, yes. Start with an audit of goals, processes, customizations and data, fix the priority problems and re-split the project into stages.

How long does an ERP implementation take?

In Deep Vision projects, typically 2–6 months; standard Sales and Inventory often go live in 6–8 weeks. Data quality, customization and team readiness matter most.

Conclusion

All seven mistakes come down to management decisions, so they can be prevented: set measurable goals, analyze processes, involve people early, limit customization, make time for training and data, and choose an experienced local partner. Clear stages help you catch mistakes while they are cheap to fix.

Want to know where the risks in your project are? Fill in a brief or call +998 77 093 00 07, and we will review your plan together in a free consultation.