"We sell to companies, it is different for us." It gets said a lot and it is nearly always used to justify not doing something. It is worth separating the real differences from the ones repeated without checking, because only the first kind should change decisions.
Starting with the false ones, which are more numerous.
Four differences that do not exist
"Selling to companies means support has to be human"
What a person who works at a company and has a problem at ten in the morning wants is what anybody wants: to fix it. The preference for solving things yourself when you can is equally well documented on both sides.
What changes is not whether they want to talk to somebody. It is that, when talking to somebody is necessary, there is less room to make them wait, because their work is stopped.
"Selling to consumers means unmanageable volume"
Volume depends on how clear the product is, not on who you sell it to. There are consumer products with fewer questions per customer than enterprise tools, simply because they explain themselves.
The metric that settles this is contact rate — questions per customer per month — and there are products of both kinds at every level of it.
"Selling to companies means more technical questions"
It means more specific questions, which is not the same. "Why did this integration stop syncing yesterday at 16:00?" is no more technical than "why is my confirmation email not arriving?"; it is more particular and it requires opening an account.
"Consumer support automates and enterprise support does not"
It is precisely backwards from what usually happens, and I get to that below.
The four that are real
1. Whoever pays is not whoever asks
This is the difference nearly all the others come out of.
When you sell to people, whoever writes in is whoever pays. Treat them badly and they cancel, and there is a direct line between the two facts.
When you sell to companies, whoever writes in is usually a user who did not decide the purchase and does not 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 lose the account anyway, or go indifferently and they renew because the sponsor never found out.
What it changes in practice: support stops being sufficient on its own. Somebody has to be watching account health — a CSM, even part time — because the signals that predict cancellation are not in the inbox. They are in the usage that drops, the sponsor who disappears and the team that stopped inviting people.
It is the difference that makes [silent churn](/en/glossary/churn) the main problem when selling to companies: the customers who leave having never written a complaint.
2. Few customers, a lot per customer
A hundred customers at €400 a month, or ten thousand at €4. The same revenue and two operations with nothing in common.
With a hundred customers, every conversation is a measurable slice of your revenue and deserves time. With ten thousand, no individual conversation moves the needle and the only thing that matters is the system.
What it changes in practice: selling to companies, you can — and should — hand-write the replies that deserve it, because the return on that half hour is enormous. Selling to consumers, half an hour on one conversation is half an hour unavailable to the other two hundred.
And a less obvious consequence: with few customers, support is a source of product information with no substitute. Ten conversations read properly say 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 is not a difference in tone, it is a difference in cost structure. Committing to a one-hour response outside office hours means having somebody on call, and that gets paid every month whether or not anybody writes.
What it changes in practice: support becomes a variable in the commercial negotiation. And a tension appears that is worth seeing coming: sales promises deadlines support has to sustain, and the conversation about whether they can be sustained has to happen before the signature, not after.
4. Which part can be resolved without a person
Here is the twist, and it runs against intuition.
Selling to consumers, a high proportion of questions are answered with general information: how something works, where an option lives, what a message means. All of that resolves with good documentation and something that can search it.
Selling to companies, more than half the questions depend on the specific account state: which plan they are on, how many seats are left, why they were charged that amount, why that integration fails only for them. A bot reading documentation resolves none of it, however good the documentation, because the article does not know who is asking.
What it changes in practice: the conclusion usually drawn here is "so AI is no use when you sell to companies", and that is the wrong conclusion. What is no use is a documentation bot. A system that can check account state before answering eats exactly the half the documentation bot cannot touch, and that half is the expensive one: the one that today makes somebody open the admin panel, check three things and come back.
Put differently: selling to consumers, automation covers a lot of cheap volume. Selling to companies it covers less volume, but every conversation it resolves was worth considerably more.
What does not change
Once all of that is separated out, there is a list of things that are identical in both cases and equally forgotten in both:
- The real cost of a ticket is the time of the person handling it, not the tool licence.
- Half of that time goes on finding context, not on writing.
- A stale article does more damage than none.
- The metric that predicts the future is contact rate per customer, not absolute ticket count.
- Hiding the contact route raises deflection rate and damages the business.
None of the five depends on who you sell to.
The question worth asking instead
Rather than "do we sell to companies or to consumers?", which decides little, there is one that decides quite a lot:
Out of a hundred real conversations, how many are answered with general information and how many require opening the account?
That split is what determines what can come off your plate and with what. And although it correlates with who you sell to, it correlates worse than people assume: there are consumer products with heavy account-specific volume — anything with a balance, orders or a subscription — and enterprise tools whose volume is almost entirely documentary.
Classifying a hundred conversations takes an afternoon and decides better than any label about the business model.
