Plugged In, Priced Up: Auditing the Real Bill Behind 'Works With Your CRM'

There's a slide in almost every live-chat demo deck that goes by without much scrutiny. It's the logo wall: HubSpot, Pipedrive, Zoho, Salesforce, Xero, a couple of helpdesks, maybe a payments processor. Sixteen colourful squares and one sentence underneath — works with your CRM.

That sentence is doing an enormous amount of unpaid labour.

For a UK SME with a sales team of two to twenty, the integration question is often the deciding factor in a purchase. Nobody wants to retype names and email addresses into a CRM at half five on a Friday. Fair enough. But "works with" is a phrase with no agreed meaning, and the gap between what vendors show in a demo and what lands in your CRM on a wet Tuesday in February is where a surprising amount of money goes missing.

This is a piece about pricing that gap.

Three different things, all called "integration"

Before you can audit the cost, you need to know which of these you've actually been sold.

Tier one: native, two-way sync. The chat platform and the CRM talk to each other directly. New contact created in chat appears in the CRM with mapped fields; a note added in the CRM can surface in the chat tool. This is the expensive, well-built version and comparatively few vendors do it properly across more than a handful of CRMs.

Tier two: one-way push through middleware. Chat tool fires a trigger into Zapier, Make or n8n, which creates or updates a CRM record. Works fine, mostly, with caveats. It's also a second subscription you didn't budget for and a second place where things can break silently.

Tier three: "no-code webhooks". This is usually marketing for we've given you an API and a documentation page. No-code is a lovely phrase. What it frequently means is that a freelancer, or your most technical colleague, will spend eight hours building something that a vendor support agent will later describe as "custom".

Tier three is not necessarily bad. It's just not free, and it is rarely quoted in the proposal.

Where the money actually leaks out

Setup fees are the obvious one — £250 to £2,000 depending on the vendor and how keen they are for the deal. Less obvious:

• Middleware subscriptions. Ten to twenty pence per task sounds trivial until you're at 900 tasks a month and the invoice is bigger than the chat seat cost itself.
• Mapping labour. Somebody has to decide that "Full name" in chat becomes "First name" and "Last name" in the CRM. Somebody has to do it again when your ops manager adds a "Territory" field in April.
• Duplicate records. Chat creates a contact. The CRM already has one from a web form, made with a slightly different email casing. Now you have two.
• Clean-up. This is the big hidden line item, and it's paid in staff hours rather than on an invoice, which is precisely why it never makes it into a business case.
• Reporting blind spots. If conversations live in chat and deals live in CRM and the two only shake hands on contact creation, your attribution is guesswork. Which of your chat conversations turned into revenue? Nobody knows.

None of that is a scandal. It's just the ordinary friction of joining two systems that were designed separately. The problem is that vendors present integration as a feature checkbox while buyers experience it as an ongoing operational cost.

A worked example: the second database

A two-person sales team, trade supplies, based just outside Leeds. They sold on the strength of fast quoting, so they bought a chat tool with a "native HubSpot integration" for a little over £90 a month plus a £450 setup fee.

The demo went beautifully. The sales director told me the word seamless came up three times.

What actually got mapped: name, email, and the chat message, dumped into a note. What didn't get mapped: company name, phone number, postcode, the product they'd asked about, or the conversation history — because the chat vendor's native sync only supported a fixed set of fields and the rest needed the API route, which needed a developer, which needed a quote the team hadn't asked for.

So the reps did what reps always do when a system doesn't serve them. They kept a spreadsheet.

Within six months they had roughly 400 chat-created contact records in HubSpot, a meaningful share of them missing a company name, a parallel spreadsheet with the details that mattered, and a growing suspicion that the CRM had stopped being a reliable place to look. Their marketing consultant spent two days a month on de-duplication. Nobody had ever put a number against that.

That's the integration tax. It's not one charge. It's a subscription you didn't know you'd bought — paid in time, in messy data, and in reps quietly losing faith in the tool.

Five questions to put to the vendor in the demo

Ask these before the trial ends, not after the invoice arrives.

1. Can you show me a field-by-field map, for our CRM, on screen? Not a list of supported apps — an actual map of which chat fields push into which CRM properties. If the answer is a PDF, ask for the CRM-side property names.

2. When we add a field next quarter, who does the mapping? If it's you, ask what that costs and how long it takes. If it's their support team, get it in writing that it's included.

3. When the sync breaks, who finds out? Silent failures are the killer. You want an alert to a named human, ideally within the hour, not a rep noticing three weeks later.

4. What's the all-in monthly cost at our seat count, including middleware? Ask them to build the number in front of you. If they can't, that's your answer.

5. Can we export every contact and every conversation as a single file, on demand, without asking anyone? This is your exit ramp and your back-up plan. If exporting is awkward, assume the data is being held slightly hostage.

One unfashionable alternative

Occasionally the honest answer is: don't integrate. If your chat volume is low and your reps are disciplined, a daily CSV export plus an agreed naming convention for new contacts will beat a half-broken sync every time. It's not impressive on a slide. It is, however, cheap, visible, and easy to fix when it goes wrong.

Integration is a genuinely good idea, and a properly built two-way sync will pay for itself many times over in a team of five. The trick is refusing to accept the logo wall as evidence of it. Ask for the map. Price the middleware. Count the clean-up hours.

The vendors who can answer all five questions without flinching are the ones worth your setup fee.