19 October 2026
In short
What does CRM integration involve for a Coventry business?
CRM integration in Coventry usually means stopping the sales system and the operational system from disagreeing. GTX Digital maps which system owns each field before writing any code, because most CRM integrations fail on ownership and duplicates rather than on the technology.
- Decide which system owns each field before anyone writes code.
- Most CRM pain is duplicates, and duplicates are a matching problem.
- One-way sync solves more than people expect and breaks far less.
- Integrations need an owner after go-live, or they rot silently.
- If sales and operations disagree about what a customer is, fix that first.
Think this applies to your business?
We will tell you straight whether it is worth doing, what it would take, and what we would do instead if it is not.
Tired of the two systems disagreeing? Book a technology review. Bring both screens and the spreadsheet somebody keeps on the side — that third one tells us the most.
Every CRM integration starts with the same complaint. The numbers in the CRM do not match the numbers in the system that does the actual work.
The technology is almost never the reason.
Who owns the field
Here is the question that decides whether an integration works.
For every piece of data that lives in both systems, which one is right when they disagree?
For each one, pick an owner. The owner can be edited. Everywhere else it is displayed and read-only.
| Field | Who should own it | What happens if both can edit |
|---|---|---|
| Customer address | Whichever system despatches the goods | This morning's correction disappears tonight |
| Credit limit | Accounts, always | Sales raises an order the business will not honour |
| Order status | The system that does the actual work | Sales quotes a date off a stale record |
| Contact phone number | The CRM | The newest number loses to the oldest sync |
Skip this and you get the classic failure: both systems can edit, the sync overwrites in whichever direction ran last, and someone's correction disappears overnight. People stop trusting it within a fortnight, and once trust goes they start keeping their own spreadsheet. Then you have three sources of truth rather than two.
This mapping takes an afternoon with the right people in the room. It saves weeks.
Duplicates, which is the real work
Ask what the biggest CRM problem is and most teams say duplicates.
The same customer exists four times. Slightly different spellings. One with a trading name, one with the limited company name, one with a typo from 2018, one created by someone who did not find the first three.
That is not a sync problem. It is a matching problem, and it has to be solved before any integration, or you will faithfully copy four customers into the other system.
Matching needs a rule, and the right rule depends on who your customers are.
| Matching rule | Works best on | Watch for |
|---|---|---|
| Company number | Anyone registered at Companies House | Sole traders and partnerships do not have one |
| Postcode plus the first few characters of the name | Most B2B records, surprisingly well | Two firms sharing one trading estate address |
| Email domain | Smaller businesses | Anyone using a generic mailbox |
Whatever rule you choose, run it over the real data and have a person review the merges. Automatic merging of near-matches is how two genuinely different customers in Nuneaton end up as one.
One direction solves more than you think
Two-way sync sounds obviously better. It is roughly four times the work and more than four times the failure modes.
Before agreeing to it, check what people actually do. Often sales need to see live order status, and nothing in the CRM needs to flow back except the new customer record.
That is one-way with a small exception, and it is a far simpler thing to build, test and keep running.
Our CRM integration work usually starts by narrowing the brief in exactly this way, and the Coventry page for it covers the local engagement.
What breaks it after go-live
Integrations are not finished when they go live. They decay.
The CRM vendor changes an API version. A field is renamed. Someone adds a mandatory field in one system and records stop saving in the other. A certificate expires. Rate limits change and the nightly run stops halfway.
So the build has to include three unglamorous things.
- Logging you can read without a developer.
- An alert when a run fails, sent to a person who will act on it.
- A way to re-run a failed batch without duplicating everything.
And it needs a named owner. Not a team — a person who gets the alert. Integrations with no owner fail quietly for weeks, and the first anyone hears is a customer asking where their order went.
The Coventry pattern
In this city the CRM is frequently the newest system in the building.
The operational software might be fifteen years old, written for the way the firm actually makes things, and entirely correct. The CRM arrived two years ago because sales wanted a pipeline.
Which means the CRM is usually the one that should give way. The old system knows what an order is. The CRM has an opinion about it.
That hierarchy is worth stating out loud, because the newer system always has the better-looking interface and the louder advocate.
For engineering businesses the awkward part is often the quote. A quote with configuration, lead time and tooling in it does not fit a standard CRM opportunity. Decide early whether the CRM holds a reference to the quote or tries to hold the quote itself. It is nearly always the former.
Which CRM you have changes the job
Not all of them are equally willing to be integrated.
The big platforms have proper APIs, decent documentation and rate limits you can plan around. The work is predictable, and the cost is mostly in your own data rather than in fighting the vendor.
The smaller sector-specific ones vary enormously. Some have a good API. Some have an API that covers forty per cent of what the interface can do. A few have nothing but a nightly export, which turns a live integration into a batch job whether you wanted one or not.
Find this out before you promise anyone a live view. A five-minute read of the vendor's developer documentation will tell you more than any sales call.
If there is genuinely no API, the honest options are a scheduled export, a database-level feed where the vendor permits it, or changing CRM. Screen-scraping a cloud product is not an option worth selling you — it breaks on their next release.
What it costs to run
Build cost gets all the attention and running cost gets none.
Expect a modest monthly figure for hosting the integration, monitoring it and dealing with the handful of failures that are nobody's fault. Expect a day or two a year when a vendor changes something and it needs adjusting.
That is not a reason to avoid integrating. It is a reason to put the running cost in the business case at the start, so the saving you claim is the real one.
A sensible sequence
Clean the data. Dedupe, with a person reviewing.
Agree the field ownership map and write it down.
Build one direction, for the handful of fields people actually need.
Watch it for a month. Add the second direction only if the month proves you need it.
If the deeper problem is that the operational system cannot expose its data at all, that is a different article: legacy software replacement in Coventry. If it is that too many people are retyping between screens, start with business automation in Coventry.
We do this across professional services and manufacturing firms, and our wider work in the city is on the Coventry page. When you want it looked at properly, get in touch.
Common questions
Why does our CRM never match the operational system?
Usually because both can edit the same fields and neither is defined as the owner, so whoever typed last wins.
Should the sync go both ways?
Only where it must. Two-way sync doubles the failure modes. Most teams need one direction for most fields.
How long does CRM integration take?
A focused one is weeks, not months — if the data is clean. The clean-up is usually the longer half.
What does it cost to run?
Budget for monitoring and the occasional fix. APIs change, and an integration nobody watches fails quietly.
Next step
Start with a technology review, not a proposal.
A short, structured conversation about where your systems are costing you time, margin or growth — and what the realistic options are. No pitch deck.