Somewhere between the tenth and the hundredth employee, every growing business collides with the software question. The system that served you faithfully is creaking; the vendor's new edition promises to fix what ails you; a rival tool demos beautifully; and someone in the family or the team has proposed building something custom. It is an expensive decision to get wrong in either direction — over-buying strains cash flow for years, and under-buying costs every day the business runs around its tools. Here is the framework we use when clients ask us to referee.
Start with the process, not the platform
The most expensive software decisions we are asked to review were made before the process was understood. A platform cannot fix a workflow you have never mapped — it will automate the confusion at a monthly subscription. So the first question is never "which software" but "what, precisely, does the work look like today, and what should it look like?" An hour of process mapping frequently changes the answer: requirements shrink, the shortlist changes, and sometimes the decision becomes "neither — the current system is fine once it is configured properly." Vendors rarely lead with that option. We do.
The three paths, honestly compared
- Buy — adopt an off-the-shelf product. Best when your process matches how the industry works, and worst when your difference from the industry is your advantage. Watch the total cost: licences compound, migration is always under-quoted, and training is never included in the demo.
- Adapt — configure and integrate what you already own. The option nobody sells you, because there is nothing to sell. Most SME platforms are substantially more capable than their owners use, and a configuration-and-integration project typically delivers a large share of the benefit of new software at a fraction of the cost and risk. This is our default first recommendation, and it is where our honesty costs us vendor friendships.
- Build — commission something custom. Reserved for processes that are genuinely a competitive advantage and cannot be bought. Custom software is a permanent commitment: it must be maintained, documented and staffed with knowledge forever. For most SMEs, most of the time, building is the wrong answer — and a custom build for a non-core process is almost always regretted.
The five questions that decide it
- Is the process stable? Never buy or build for a workflow that is still changing — you will pay to automate a draft.
- How different is your process from the industry's? The more standard it is, the better "buy" fits; the more it is your advantage, the more carefully you should think before forcing it into someone else's template.
- What is the three-year total cost? Licences, migration, training, integrations, and the internal effort of administration — modelled honestly.
- Who does the work of adoption? Every path needs a champion and protected training time. A system nobody adopts is a donation to its vendor.
- What is the exit? Can you get your data out? Under what terms? A tool you cannot leave is a tool that can raise its prices.
Our record on this question
Over thirteen years we have recommended against purchases more often than for them — and those conversations, uncomfortable as they can be in the moment, are why clients keep asking. A configuration project delivered last year the value a new platform promised next year, at a third of the cost, is not a rare story in our practice; it is close to the median. The software industry is built on the assumption that every problem needs new software. Our practice is built on the assumption that your problem needs the smallest effective change — and sometimes that is new software, sometimes it is configuration, and sometimes it is a process fix that costs nothing but attention.
If you are facing this decision, bring us the shortlist and the process map. If you do not have the process map, that is the first engagement — and it will serve you long after the software question is settled.
A referee without a horse in the race
We take no vendor commissions — our recommendation is clean by construction.