"We sell to companies, so it's different for us." People say this a lot, and they usually say it to justify not doing something. It helps to separate the differences that hold up from the ones repeated without checking, because only the first kind should change your decisions.
The false ones come first, because there are more of them.
Four differences that don't exist
"Selling to companies means support has to be human"
Someone at a company with a problem at ten in the morning wants what anybody wants: to get it fixed. The preference for solving things yourself when you can is just as well documented on both sides.
The difference shows up when a person is needed. Business customers have less patience for waiting, because their work has stopped.
"Selling to consumers means unmanageable volume"
Volume depends on how clear the product is. Some consumer products get fewer questions per customer than enterprise tools, simply because they explain themselves.
The metric that settles it is contact rate, the number of questions per customer per month. There are products of both kinds at every level of it.
"Selling to companies means more technical questions"
It means more specific questions. "Why did this integration stop syncing yesterday at 16:00?" is no more technical than "why isn't my confirmation email arriving?". It is more particular, and answering it means opening the account.
"Consumer support can be automated and enterprise support can't"
It usually works the other way round. More on that below.
The four differences that hold up
1. The person who pays isn't the person who asks
Almost every other difference comes from this one.
When you sell to people, whoever writes in is whoever pays. Treat them badly and they cancel, and the line between the two is direct.
When you sell to companies, whoever writes in is usually a user who didn't make the purchase and won't decide the renewal. The person who signed may not have opened the product in eleven months. That breaks the loop: support can go well and you still lose the account, or go so-so and they renew because the sponsor never found out.
In practice, support stops being enough on its own. Somebody has to watch the health of each account, a CSM even if only part time, because the signals that predict cancellation aren't in the inbox. They are in usage that drops, a sponsor who goes quiet and a team that stops inviting people.
This is why [silent churn](/en/glossary/churn) is the main problem when you sell to companies: customers who leave without ever writing a complaint.
2. Few customers, a lot per customer
A hundred customers at €400 a month, or ten thousand at €4. Same revenue, two operations with nothing in common.
With a hundred customers, every conversation is a visible slice of your revenue and deserves time. With ten thousand, no single conversation makes a difference, and the only thing that matters is the system.
In practice: when you sell to companies, you can and should hand-write the replies that deserve it, because that half hour pays off many times over. When you sell to consumers, half an hour on one conversation is half an hour the other two hundred don't get.
There is a less obvious consequence. With few customers, support is a source of product insight that nothing else replaces. Ten conversations read carefully tell you more than an analytics dashboard.
3. What gets signed
Contracts with companies bring service level agreements, committed hours, sometimes availability clauses and nearly always a security questionnaire before signing.
That changes your cost structure. Committing to a one-hour response outside office hours means having somebody on call, and you pay for that every month whether anybody writes or not.
In practice, support becomes part of the commercial negotiation. Expect some tension: sales promises response times that support has to meet, so the conversation about whether they can be met has to happen before the contract is signed.
4. How much can be resolved without a person
This is the counterintuitive one.
When you sell to consumers, many questions are answered with general information: how something works, where an option lives, what a message means. Good documentation and something that can search it handle all of that.
When you sell to companies, more than half the questions depend on the state of that specific account: which plan they are on, how many seats are left, why they were charged that amount, why an integration fails only for them. A bot that reads documentation answers none of these, however good the documentation is, because the article doesn't know who is asking.
People often conclude from this that AI is no use when you sell to companies. A documentation-only bot isn't. A system that can check the account before answering handles exactly the half the documentation bot can't touch, and that half is the expensive one: today it makes somebody open the admin panel, check three things and come back.
Put another way: with consumers, automation covers a lot of cheap volume. With companies, it covers less volume, but each conversation it resolves is worth considerably more.
What doesn't change
Once all of that is sorted, a few things are identical in both cases, and both tend to forget them:
- What a ticket costs is the time of the person handling it, much more than the tool licence.
- Half of that time goes on finding information, not on writing the reply.
- An outdated article does more damage than no article.
- The metric that predicts the future is contact rate per customer, not the total number of tickets.
- Hiding the way to contact you lifts the deflection rate and hurts the business.
None of the five depends on who you sell to.
A better question to ask
"Do we sell to companies or to consumers?" decides very little. This question decides a lot:
Out of a hundred conversations from your inbox, how many can be answered with general information and how many require opening the account?
That split tells you what can come off your plate and with which tool. It does correlate with who you sell to, but less than people assume. Some consumer products have a lot of account-specific volume (anything with a balance, orders or a subscription), and some enterprise tools get almost only documentation questions.
Classifying a hundred conversations takes an afternoon, and it tells you more than any label about your business model.
Frequently asked questions
Does B2B support have to be 24/7?
Only if you signed for it. Business customers don't mind that you answer at nine; they mind not having been told you answer at nine. And if an agent handles account questions at any hour, only what needs a person waits for office hours.
Do I need a CSM if I sell to companies?
From a few dozen accounts on, someone has to watch the health of each one, even part time, because B2B churn doesn't announce itself in the support inbox.
Can B2B support be automated?
The half that depends on the account, yes, with an agent that checks it. The half that needs someone's judgment, no, and that is the most valuable part. A documentation bot automates little in B2B, because almost nothing these customers ask is in an article.
What if I sell to both?
Measure the split separately. The consumer segment usually turns out to be mostly documentation questions and cheap to automate, and the business segment depends on the account and is worth both the agent and the person.
What SLA can I promise?
The one you already meet in 95% of cases with your current team, measured over three months. Promising before measuring is the fastest way to break it.
