Answer
How do you choose business software?
By running your real work through it for a fortnight, not by comparing feature lists nobody will use.
Run a fortnight of real work through the shortlist. Feature comparisons select for breadth and the failures are always workflow-shaped — the awkward case, the missing integration, the step that takes six clicks.
Software gets selected on feature comparison and fails on workflow, which is why the two most common outcomes are a product that does everything and is not used, and a product chosen for a feature nobody ended up needing. A demonstration shows the intended path, walked by someone who knows it. The awkward cases — where your work does not match the assumed shape — are where the product will actually be judged, and they never appear in the demonstration.
The reliable test is to run real work through candidates for a fortnight. Take the enquiries, jobs or records you actually processed, put them through each shortlisted product exactly as you would in practice, and note where it becomes slow or impossible. Two weeks of that produces a better decision than any amount of comparison, and it surfaces the specific mismatches that make a product unusable in a business like yours.
Integration is the criterion most underweighted at selection and most decisive afterwards. A product that cannot exchange data with the systems you already run turns into a second place where information lives, which is worse than the problem it was bought to solve. The question to ask is not whether it integrates but with which specific products, in which versions, in which direction, and what happens when the other side changes.
Data portability decides how much leverage you retain and it should be established before signing rather than at renewal. Can you export everything you put in, in a usable form, including the history and the attachments? A supplier who cannot answer that clearly is selling something you can only leave by starting again, which is a commercial position rather than a technical limitation and should be priced accordingly.
The people who will use it should be part of the selection, and not as a courtesy. They will identify in an hour what a decision-maker will not notice for a quarter — that a routine action takes six clicks, that the mobile version drops a field, that the search does not find things the way they look for them. A product chosen without them is a product with an adoption problem already built in.
Cost comparison needs the same discipline as any other purchase here: licence is one line among four. Migration of existing data, configuration, training, and the ongoing ownership of keeping it correct all have real costs, and the cheapest licence with the most expensive migration is common. A product that fits your workflow at a higher price is nearly always cheaper in total than one that requires the business to change shape.
Every product wins its own demonstration, because the demonstration is the one path the product was built to walk.
Answer Production Engine, Context Theory
Related questions
Should we choose the market leader to be safe?
It reduces some risks and creates others. Market leaders are stable and well documented, and they are frequently built for larger organisations, which shows up as complexity a small business pays for and does not use. Safety in this decision comes from fitting your workflow and being able to leave, not from the supplier's size.
How long should we expect a system to last?
Long enough to justify the migration cost, which for a small business usually means several years, and that expectation should shape the portability question rather than the feature question. Software you can leave cleanly can be chosen quickly; software you cannot leave has to be chosen carefully, because the decision is effectively permanent.
METHOD
Every figure below carries its source and the date it was verified. Nothing on this page is asserted.
The numbers on this page.
| What | Value | Specific to |
|---|---|---|
| US SMB retainer, focused one-to-two-service engagement | $1,500–$4,000 | Category-wide |
| Meaningful market band | $1,500–$10,000 | Category-wide |
| Firms that never responded to a web enquiry at all | 23% | Category-wide |
2026 agency pricing survey · per month · verified
Oldroyd, McElheran & Elkington, "The Short Life of Online Sales Leads", Harvard Business Review (March 2011) · 1.25M inbound leads across 2,241 US firms · verified
What is specific to this page.
| Kind | Claim | Check it against |
|---|---|---|
| Procurement | A demonstration walks the one path a product was built for, so the awkward cases where a business's work does not match the assumed shape — which is where the product is actually judged — never appear in it. | Running the business's least standard recent job through the candidate product rather than a typical one. |
| Software | A product that cannot exchange data with existing systems becomes a second location for information, which is worse than the problem it was bought to solve, so the integration question must name specific products, versions and direction. | The supplier's published integration list, checked against the specific versions the business runs. |
| Procurement | Inability to export all data including history and attachments in a usable form is a commercial position rather than a technical limitation, and it determines the buyer's leverage at every renewal. | The supplier's contractual export provisions and the formats it commits to on termination. |
| Workflow | The people who will operate a system identify workflow failures within an hour that a decision-maker will not notice for a quarter, such as routine actions requiring many steps or search behaving unlike how they look for things. | Having intended users complete their three most frequent tasks in each candidate and timing them. |
Each row would be wrong on another industry's page. Where a sourced figure exists it is in the table above instead; these are the constraints that shape the work and do not happen to be numbers.
Start with the measurement.
Reading about a benchmark is not the same as knowing your own number. The audit produces yours, measured rather than estimated.
$497 · delivered in 5 business days · credited against month one