Failed Projects, Contracts and Official Sources
Why ERP projects fail in Indian MSMEs, and how to avoid the usual patterns
The recurring patterns behind ERP overruns and disappointments, early warning signs, what each costs and how to prevent it, described generically without naming companies.
ERP projects rarely fail because the software cannot work. They fail, or disappoint, because of a handful of recurring patterns: scope that was never written down, data migration left until the last weeks, customisation that breaks at every upgrade, key users with no time for the project, choosing on headline price instead of total cost, and contracts that make leaving expensive. Survey figures on how many projects overrun differ a great deal between studies and are hard to compare, so this guide does not quote one; the patterns themselves are consistent. Each is described below with its warning signs, typical cost and prevention. Examples are generic and illustrative; no real company is described.
This is the overview for the failed projects, contracts and sources section, and links to every guide in it.
Pattern 1: unclear scope
What happens. Requirements are discussed in meetings but never written down and signed off. During build, new needs appear; the partner raises change requests; the budget grows and go-live slips.
Warning signs. No solution design document; "we will configure as we go"; disagreements about whether something was "included".
Prevention. A written requirements list (checklist), a signed solution design, and a change-request process with prices and approvals.
Pattern 2: data migration left late
What happens. Masters with duplicate customers, wrong GSTINs and inconsistent item names are loaded at the last minute. Opening stock and balances do not reconcile. GST returns from the new system disagree with the old books.
Warning signs. No data owner; the first test load is scheduled close to go-live; "we will clean it in the new system".
Prevention. Start cleaning on day one, run at least two test loads, and have your accountant sign off reconciliations. See implementation and migration costs.
Pattern 3: customisation creep
What happens. The business asks the ERP to replicate every old report and habit. Custom code accumulates. Each vendor update risks breaking it, and upgrades are postponed, which can leave statutory updates (e-invoice schemas, return formats) blocked.
Warning signs. Many customisation line items; no one asking "could we change the process instead?"; custom code owned by one developer.
Prevention. Default to standard features; justify each customisation by its business value; agree in writing who maintains it through upgrades and at what cost. See hidden ERP costs.
Pattern 4: no time from key users
What happens. The people who know the processes are also running daily operations. They miss workshops and testing, so the configuration reflects the partner's assumptions. After go-live, users find workarounds or go back to spreadsheets.
Prevention. Name key users, plan their hours, backfill routine work during peak weeks, and make the owner or finance head the project sponsor with a weekly review.
Pattern 5: choosing on headline price
What happens. The lowest per-user price wins. Later the business discovers paid add-ons for e-way bills or multi-GSTIN use, a minimum user count, a weak partner, or a high uplift at the first renewal.
Prevention. Compare normalised five-year TCO, not list prices. See comparing ERP quotes and the TCO estimator.
Pattern 6: contracts that trap you
What happens. The contract auto-renews with a short notice window, allows uplifts to "then-current list price", and gives only partial data export after termination. Leaving becomes expensive just when the relationship is worst.
Prevention. Negotiate exit, data export, uplift caps and SLA before signing. See the SaaS contract guide.
Pattern 7: compliance treated as an afterthought
What happens. Go-live succeeds, but the audit trail was switched off during migration, back-ups are stored outside India for a company, or employee data is accessible to everyone. These come to light at the statutory audit or after an incident.
Prevention. Make compliance part of acceptance testing: see the audit trail guide and the data protection and CERT-In guide.
What each pattern typically costs (illustrative)
| Pattern | Where the cost shows up | Illustrative impact on a ₹40 lakh five-year budget |
|---|---|---|
| Unclear scope | Change requests, delay | +10–25% of implementation |
| Late migration | Rework, wrong returns, delayed close | Several weeks of staff time; possible GST corrections |
| Customisation creep | Upgrade cost, blocked updates | Recurring cost every upgrade |
| No key-user time | Low adoption, workarounds | Savings in the business case never arrive |
| Headline price | Add-ons, uplifts | +5–20% of recurring cost |
| Contract trap | Exit cost, forced renewal | One extra year of fees, migration cost |
| Compliance afterthought | Audit qualification, incident costs | Hard to quantify; reputational |
These are illustrations of scale, not statistics.
A pre-signing health check
Before signing, you should be able to answer yes to each:
- Is there a written, prioritised requirements list?
- Have we seen demos on our own data for our key scenarios?
- Do we have a normalised five-year TCO for at least two options?
- Are implementation scope, exclusions and change-request rates in writing?
- Do we have named key users with planned time?
- Does the contract cap uplifts and guarantee full data export on exit?
- Have compliance requirements (e-invoicing, audit trail, back-ups, data protection) been tested, not just promised?
- Have we budgeted only grants that are sanctioned? See government support for MSME digitisation.
Use the ERP selection and contract checklist to track the answers, and the official sources list to verify rules.
FAQ
What should I do if my project is already overrunning?
Stop and re-plan: freeze new scope, list what is needed for a minimum viable go-live, re-estimate with the partner, and decide which items move to a second phase. Escalate formally under the contract if the partner is not performing.
Is it better to switch partner mid-project?
Sometimes, but it costs time and knowledge. Check who owns the configuration and custom code, and whether the software vendor can recommend another partner, before deciding.
Can I get compensation if a project fails?
That depends on your contract and the facts. For significant disputes, take legal advice early and keep records of scope, approvals and delays.
Sources
- Ministry of Corporate Affairs — audit trail and books of account rules
- CERT-In — incident reporting directions
- MeitY — data protection framework
- CBIC GST — compliance obligations the system must meet
- Ministry of MSME — MSME support schemes
