← Insights Say hi πŸ‘‹
Sales systems Β· Plain-English guide

Off-the-shelf, or built for you?

The short answer

Buy off the shelf when your process is simple, settled or still being invented, and when nobody internally has time to own a system. Have one built when your process is genuinely your own, the team is big enough that small frictions compound, and someone will say what should change each month. The real question is not price. It is which one bends, the software, or the way your team works.

The buy-or-build question is usually argued as a money question. It is not. A ready-made CRM and a built one both cost real money, and for most businesses the ready-made one is cheaper by a distance.

The question worth asking is quieter than that: when your process and the software disagree, which one gives way? Everything else follows from the answer.

What you are actually buying

A general CRM is the average of ten thousand companies. It has to work for a recruitment desk, a manufacturer, a design agency and an insurance team, so it is built around the parts they all share, contacts, deals, stages, notes, a dashboard.

That averaging is not a flaw. It is the reason the thing costs a few hundred dollars a month instead of a few hundred thousand, and why it arrives working on day one instead of in four months. You are buying the accumulated opinions of a company that has watched thousands of businesses sell things. Often those opinions are better than yours.

The cost of the average is that it fits nobody exactly. So every team that buys one does the same thing: they reshape themselves to fit it. Slightly. Quietly. Usually without deciding to.

What configuration can fix, and what it cannot

Most ready-made systems can be configured a long way, and people underrate how far. If your disagreement with the software is about naming, ordering or which fields appear, configuration will almost certainly solve it, and you should stop reading and go and configure it.

Configuration handles this comfortably:

Where configuration stops is more specific than "the product isn't flexible enough". It stops when your process needs a concept the product has no idea exists.

A recruitment desk, for example, has two sides, a candidate and a job, and the interesting object is the pairing between them, which has its own life and its own history. A product built around one customer moving through one pipeline can be bent to fake that, but it does not believe in it. A financial adviser has a household, not a contact: a husband, a wife, three policies and a review date that belongs to all of them. A field service business has the same customer at four different addresses with four different histories.

In every case you can force it. You will end up with duplicate records, a field called "linked to", and a rule everyone has to remember. That is the moment the decision actually arrives, and it is worth noticing when it does, because from here the costs stop being visible on the invoice.

The hidden cost of a workaround

A workaround is not free and it is not stable. It fails in a particular sequence, and once you have seen the sequence you can spot it early.

First it costs time, thirty seconds of remembering, per record, per person, forever. Thirty seconds does not sound like anything. Multiply it by twelve people entering forty records a week and it is a working day a month spent on a rule that exists only because the software does not understand your business.

Then it costs accuracy. A rule that lives in people's heads is followed unevenly. Some people use the linked field, some do not, and the reports built on top of it are now quietly wrong in a way that is very hard to see from a dashboard.

Then it costs the record itself. This is the part that matters. When something is awkward to enter, people stop entering it, not the whole record, just the awkward part. The nuance goes into a note, or a chat message, or nowhere. And the moment a piece of your process stops being recorded, someone rebuilds it somewhere the software cannot object to.

That somewhere is a spreadsheet. It is always a spreadsheet.

The spreadsheet is the diagnosis, not the disease. When a team keeps a sheet alongside the CRM, it is not laziness, it is a precise map of everything the system will not let them record. Ask to see it before you decide anything. If the columns are things your CRM could hold with an afternoon of configuration, you have a configuration problem. If the columns describe a way of working the product has no concept of, you have found the real question.

When off-the-shelf is the right answer

Often. Probably more often than not, and it would be dishonest to write this page without saying so plainly.

Your team is small and the process is simple. Three people and a short sales cycle do not need a bespoke system. They need somewhere shared to put things, and almost anything does that. At that size the bottleneck is not the software, it is that there are three of you. Buy something, spend a week setting it up properly, and go and sell.

You have not figured out your process yet. This is the strongest reason of all and the most commonly ignored. Building a system around a process you are still inventing bakes today's guesses into software, and software is much harder to change your mind about than a habit. A generic product is a useful constraint while the shape of the work is still moving, it forces you to articulate your stages in someone else's vocabulary, which is a good exercise. Come back when you can describe how you sell in ten sentences and mean all of them.

Nobody internally will own it. A built system needs a person who decides what changes next month. Not a committee, not "the team", a person. If you cannot name them, the system ships, ages, and within a year you are back to reshaping yourselves to fit software, except now it is software nobody else maintains. That is the worst of both.

What you have is genuinely working. If your team opens the CRM without being asked, the pipeline in it is true, and nobody keeps a private sheet, you are finished. There is no upside worth the disruption. Improving something that works is a much worse use of a year than fixing something that does not.

Your process is normal and you are not precious about it. Some businesses sell in a way that is genuinely ordinary. Enquiry, quote, follow up, close. If that is you, the average of ten thousand companies is a very good description of your company, and you should take it gratefully.

The case for building only gets strong when three things are true at once: the process is genuinely your own, the team is large enough that thirty-second frictions compound into real money, and there is someone who cares enough to keep shaping it. Two out of three is not enough.

What "built" should mean

The honest objection to building is that it is finished the day it ships, and then it ages while the bought one keeps getting updates. That is a fair description of most custom software, and it is why so much of it is a disappointment.

A one-off project is a snapshot of how you worked during the four months someone was writing it. Your business does not stop during those four months. Neither does it stop afterwards. Within a year the system describes a company you used to be, nobody wants to pay to change it, and people start working around it, which is exactly the problem you built it to escape, arrived at from the other direction.

So the only version of "built" worth buying is one that keeps changing. In practice that is unglamorous and looks like this: someone asks for a field and it appears that week. A stage turns out to really be two stages. A report nobody has opened in three months gets deleted. Over a year that is dozens of small corrections, each one moving the system closer to how the work is actually done rather than further away.

Which means the question to ask anyone offering to build you something is not what it will do. It is what happens in month eight, who changes it, how quickly, and what that costs. If the answer is a new quotation each time, you are buying a snapshot.

Someone has to own it

The most common reason a built system disappoints has nothing to do with the software. It is that nobody inside the business was responsible for it.

The owner does not need to be technical. They need to be close enough to the selling to know what is annoying, and senior enough that when they say "we are doing it this way now", it happens. Usually that is a sales manager or the founder, not IT.

Their job is small but constant: collect what the team complains about, decide which three things matter this month, say no to the rest, and make sure the changes actually get used. An hour or two a month. Without that hour the system drifts, because nobody else is looking at it from the inside.

If reading that paragraph produced no name, that is a genuine answer to the buy-or-build question, and a useful one to get for free.

Migration and lock-in, on both sides

People raise lock-in as an argument against building. It is a fair concern and it is worth being clear that it cuts both ways.

With a ready-made product you are locked in by habit and data shape. Your information is arranged the way that product arranges it, your integrations are built against it, your team has learned its screens, and the price is whatever it becomes at renewal. Export is usually possible and usually painful, you can get the rows out, but not the automations, the layouts or the years of small configuration decisions.

With a built system you are locked in by people. The risk is not the data, it is whether anyone other than the original builder can work on it. That is the question to ask up front, and the answers should be boring and specific: the code is in a repository you own, the data sits in a standard database you can point another developer at, and there is an export that produces files you can actually open.

Either way, ask the same question before you start rather than after: what exactly would I be handed on the last day? Vague answers are the answer.

One practical note on migration, because it catches people out. Moving into a new system is mostly not a technical job. Moving the rows across is a week. Deciding which of eleven thousand contacts are real, which duplicates to keep, and what the eight legacy stages map to, that is the actual work, and no software does it for you. It is worth doing regardless. Most teams find the exercise more useful than the migration.

How to decide in an afternoon

Sit with two or three of the people who use the system every day and ask three questions.

Then check the third condition honestly: is there a person who will own it. If not, the answer is off-the-shelf regardless of the first two.

The short version

Off-the-shelf software is the average of ten thousand companies, which is exactly why it is affordable, fast to start and usually right. The price of the average is that you bend to fit it, and the parts of your process that will not bend stop being recorded.

Building is worth it when what will not bend is the part that makes you money, and only if someone keeps shaping it afterwards. A system that stops changing starts drifting away from the work on the day it is finished, whoever wrote it.

Common questions

Almost always upfront, and the comparison people make is usually the wrong one. A monthly licence is not the full cost of buying, add the configuration, the integrations, the workarounds and the time your team spends feeding something that does not fit. Compare total cost over three years rather than the invoice, and be honest that for a small team the bought one still usually wins.

Very often, yes, and you should try that first. Configuration handles naming, stages, extra fields, simple automation and reporting, which covers most complaints. It runs out when your process needs a concept the product has no idea exists, at that point you are choosing between an awkward workaround and changing how you work.

The first useful version should be weeks rather than months, because a system nobody has used yet is a set of assumptions. Build the one thing the team already wishes it had, put it in their hands, and let what they say next set the direction. Anyone quoting a six-month build before a single person has touched it is guessing for half a year.

Settle this in writing before you start, not afterwards. You should own the code and the database, and be able to hand both to another developer without asking permission. Ask what exactly you would be handed on the last day, if the answer is vague, treat that as the answer.

One named person, close to the selling and senior enough to make a decision stick, usually a sales manager or the founder rather than IT. Their job is an hour or two a month: gather the complaints, pick the few that matter, and make sure the changes get used. If you cannot name that person, buy something instead.

It is a symptom, and a useful one. The columns in that sheet are a precise list of everything the system will not let them record. If those columns could live in your CRM with an afternoon of setup, fix the setup. If they describe a way of working the product has no concept of, you have found the real question.

Eugene
Eugene

We build websites and run SEO & AEO for Singapore businesses. HeyAda is a Singapore web studio, design, copywriting, code and SEO, handled end to end. We write these guides to share what actually works, in plain English. If your site isn't getting found, say hi.

Not sure which side you are on?

Tell me how your team sells, what they keep outside the system, and how many of them there are. I will tell you whether this is worth building or whether you should configure what you already have, and I will say so plainly if off-the-shelf is the better answer for you.