Why SMBs get sold enterprise CRMs, and lose.

A multi-unit franchise group I worked with bought an enterprise CRM in the spring. Nine locations, about forty people total, two of whom sold anything. The contract was signed on a Thursday. There was a kickoff deck. There was a named implementation partner with a headshot.

Fourteen months later, the general manager was still running the actual pipeline out of a Google Sheet named LEADS_v4_FINAL. The CRM existed. People logged into it the way you visit a relative you do not like: rarely, briefly, and only when someone asks whether you have.

Nobody was lazy. Nobody was stupid. They had bought a system that assumed a job nobody at that company had.

The demo is not the product

Here is the mechanic. Enterprise CRM demos are performed on a sandbox that has been groomed for years: custom objects mapped to a real sales process, clean historical data, dashboards built by a person whose entire job was building dashboards, automations wired by an admin who knew the schema cold.

What you sign for is the same software with none of that. An empty box with a very good sales team attached to it. The gap between the demo and the box is not a licensing gap. It is a labor gap, and it gets handed to you at closing without anyone naming it out loud.

Gartner has been citing CRM implementation failure rates in the thirty to seventy percent range for the better part of two decades, depending on how failure is defined. The number bounces around because the definition does, but the direction has never moved. These projects fail at a rate that would end any other category of purchase. They do not, because the failure is quiet and it looks like a people problem.

Seat price is the smallest number in the deal

Operators evaluate CRMs by comparing per-seat pricing, because that is the number on the pricing page. It is roughly the least important figure in the entire decision.

Mid-tier and upper-tier editions from the major enterprise vendors commonly list well north of a hundred dollars per seat per month before add-ons, and the features operators assume are included (real automation, decent reporting, API access at a usable rate limit) tend to sit one tier above whatever you were quoted. That is not a trick. It is the standard packaging model, and it works because nobody knows which tier they need until they are already in.

Then comes implementation. Common industry guidance puts enterprise CRM implementation services somewhere between one and three times first-year license cost. For a company with a real sales operations function, that is a normal line item. For a nine-location franchise group, it is the entire technology budget spent before a single lead has been logged.

And the largest cost never appears on any invoice: the hours your operations manager spends becoming an amateur systems administrator instead of running operations.

The role you did not know you were hiring for

Enterprise CRMs are not products in the way a phone is a product. They are platforms. Platforms assume an owner.

Somebody has to decide what a "qualified" lead is and enforce that definition in a picklist. Somebody has to notice when three reps invented three different stage names. Somebody has to fix the automation that fires at 2am and emails four hundred people. At a company of eight hundred, that person exists, has a title, and gets paid for it. At a company of forty, that person is whoever was in the room when the vendor asked "who wants admin access."

That is the actual purchase. You bought a platform and, without a requisition, a part-time job. In our experience that job runs somewhere between five and ten hours a week during the first year, and it lands on the person who was already the most overloaded, because they were the one who cared enough to raise their hand.

Four ways it comes apart

1. Adoption decays quietly

It never dies on a Tuesday. Week one is enthusiastic. By month three, one rep is updating deals only before the pipeline meeting. By month six, the real pipeline has migrated back to a spreadsheet or a text thread, and the CRM has become a reporting chore performed for management. Sales leaders have complained about seller adoption in survey after survey for twenty years running. The pattern is durable enough that you should plan for it rather than hope your team is different.

2. Field sprawl

Because you can add a field, someone does. Then someone else adds a nearly identical one. Eighteen months in there are four fields that all mean "how did they find us" and none of them is reliably populated, which means the reporting built on top of them is decorative.

3. The integration tax

The CRM has to talk to the booking system, the phone system, the accounting package, and whatever the franchisor mandates. Each connection is a small project, and every one becomes a permanent maintenance obligation. Nobody budgets for the fifth Zapier zap. Everybody eventually pays for it.

4. Reporting theater

The dashboards get built, they look excellent, and no decision changes as a result. When the underlying data is entered inconsistently by tired people at the end of the day, a beautiful chart is just a confident-looking average of guesses.

Three questions before the next demo

We are not anti-CRM. Nearly every operator we work with needs a system of record. The question is what shape it should be, and these three tend to settle it.

Who owns this on day ninety? Name the human. Give them the hours in writing. If you cannot name them, or if the honest answer is "we will figure it out," you are not ready to buy a platform. You are ready to buy a tool.

What is the smallest thing that fixes the pain? Most SMB CRM pain is one of three things: leads getting lost, no visibility into what is open, and follow-up that depends on somebody remembering. Those are solvable with a shared pipeline, a routing rule, and a reminder. That is a fraction of a platform, and it can be running this month.

What happens if we leave? Ask, in the room, whether you can export every record, every note, every attachment, and every custom field into a format you can actually use. Ask who owns the data if you stop paying. Get the answer before you sign, not during the renewal negotiation when leverage has quietly changed hands.

Sized to the process you actually have

The alternative is unglamorous and it works: build the system around the process that already exists, rather than buying a process and hoping people adopt it.

For that franchise group, what eventually stuck was not a platform. It was a lightweight pipeline that mirrored the six stages they already used verbally, a form on the site that routed leads by location, an automatic text to the manager on duty, and one weekly report that fit on a phone screen. Roughly two weeks of work. Adoption was near total inside a month, for the boring reason that it did not ask anyone to work differently.

It also cost less annually than the previous seat licenses, but that was not the win. The win was that the data lived in a database they controlled, in a schema that matched how they talked about the business, and it could be exported in full at any time by someone who did not need a certification to do it. When we walked out, the tool did not leave with us.

When enterprise is the right answer

To be fair to the category, sometimes it is. If you have more than roughly fifteen quota-carrying sellers, a dedicated operations person, compliance requirements that demand audit trails, or a franchisor mandating a specific platform, buy the enterprise system and staff it properly. The tooling is genuinely good at that scale. It is built for the shape of company you have.

The failure is not enterprise software. The failure is buying the org chart of a company you do not have yet, then spending two years of operator attention pretending you do.

Aspiration is a fine reason to write a plan. It is an expensive reason to sign a contract.