# ERP data quality: what to fix to start a TMS

[Source](https://onechain-tms.com/en/blog/qualite-des-donnees-demarrer-un-tms/)

Language: en

Author: OneChain

Published: 2026-11-18

Updated: 2026-11-19

**You do not need to clean all your ERP data first. Identify the data the first scope’s use cases depend on and the errors that would stop them, start with the reliable fields, then set up a routine for fixing issues in the ERP. A big upfront clean-up takes a lot of work and does not guarantee the quality of future data.**

This article sets out the approach for a TMS project through three examples: addresses, postal codes and weights.

## Why does a TMS project expose ERP data issues?

Because it uses the data in a new way. A rough address goes unnoticed as long as a driver knows the customer. It becomes a problem when software has to apply a rate card or build a transport order from it.

These issues are usually already in the ERP. They surface at one of three points: during a demo with your own data, during design, or after go-live. The last is the worst case, because errors then affect real shipments. To catch them early, run demos and design workshops on an extract of your data.

## Which data can block transport operations?

Addresses and product data are the two areas to check first. Your scope may add others.

For addresses, three issues come up: duplicate records, incomplete addresses and missing postal codes. They get in the way of applying rate cards, executing orders and matching deliveries to transport. A missing postal code can make it impossible to identify the right rate card when rates are set by zone. A partial address complicates pickup or delivery.

For products, the typical example is weight. Booking transport requires an estimate of the shipment’s gross weight, and that calculation relies in part on the net weight of the items. If net weight is missing or unreliable, so is the estimate.

The checks you need depend on the use cases you choose. A first scope focused on tracking does not have the same requirements as one that applies rate cards or prepares bookings. No single field list fits every project.

## Should you clean everything before you start?

No. A full database clean-up is a heavy workload, and it does not deal with where errors come from: new entries will repeat the same issues unless something changes at the source.

Look at the data as it actually is:

1. List the use cases in the first scope, such as costing, booking or tracking.
2. For each one, note the fields it cannot work without.
3. Check those fields on a real extract: missing values, duplicates, inconsistent values.
4. Start with the reliable fields and account for the limits of the rest.

Standardized fields can be a useful anchor. The Incoterm, the rule that splits costs and risks between seller and buyer, takes its values from a fixed list. Still check that it is filled in and correct in your context: a standardized field can be filled in wrongly.

Starting with imperfect data does not mean accepting errors that block the chosen use case. If postal codes are missing for the first scope’s destinations and rates depend on them, fix them before launch or pick a different scope. The guide [TMS implementation: pilot first or big bang?](https://onechain-tms.com/en/blog/deployer-un-tms-premier-perimetre/) helps with that choice.

## Who fixes the data, and where?

A named person or team, in the ERP. The TMS shows what is happening in operations and brings issues to light; it does not replace fixing them at the source. An address corrected only in the TMS will come back wrong with the next order.

Set up a simple routine:

- users report issues they spot during operations;
- the named owner collects these reports and sets priorities;
- records are corrected in the ERP by the people who have the rights to do so;
- the routine continues with each new module rolled out, since each one draws on different data.

This split of responsibilities is agreed in the project. On its [integrations page](https://onechain-tms.com/en/integrations/), OneChain states that each field has an agreed source of truth and that the project defines who detects an issue and who corrects the data. The [deployment and support page](https://onechain-tms.com/en/deployment-and-support/) devotes one step of its method to preparing data and interfaces, including the sources to bring in and the checks that matter.

## How do you track the data that needs fixing?

With a table built around the first scope’s use cases, field by field. The one below uses the examples from this article; treat it as a template to extend with your own use cases, not a complete list.

| Data | Use in the TMS | Consequence of an error | Who fixes it |
|---|---|---|---|
| Delivery postal code | Identify the applicable rate card | Rate card not identified, cost cannot be calculated | Named owner, customer record in the ERP |
| Full address | Execute the transport order | Pickup or delivery complicated | Named owner, customer record in the ERP |
| Duplicate addresses | Link deliveries and transport | Deliveries and transport hard to match | Named owner, customer record in the ERP |
| Item net weight | Estimate gross weight for booking | Booking based on the wrong weight | Named owner, item record in the ERP |

If applying rate cards is part of the first scope, see the [rate management page](https://onechain-tms.com/en/rate-management/). If your ERP is changing at the same time, the guide [ERP migration: should you wait before launching a TMS?](https://onechain-tms.com/en/blog/migration-erp-et-projet-tms/) helps sequence both projects. The other scoping decisions are brought together in [TMS project: eight questions to settle before you start](https://onechain-tms.com/en/blog/cadrer-un-projet-tms/).

To prepare your first scope, [let’s talk about the data it depends on](https://onechain-tms.com/en/contact/?sujet=Data%20and%20TMS%20project). Describe the use cases you have in mind and the issues you already know about; do not send data extracts or confidential documents through the form.

