Consulting is worth buying when you have a decision to make and not enough information or experience to make it. It is worth very little when it is bought as a substitute for deciding, which is how these engagements end in a document that changes nothing and a vague sense that it was expensive.
Buy software consulting services against a written question
Write down the question before engaging anyone: should we build or buy this, why does our release process take three weeks, is this codebase worth keeping, what should we do about the system nobody understands. Then agree what an answer looks like. Engagements scoped around a question end; engagements scoped around an area continue indefinitely, and the difference is visible in the first meeting.
When a software developer consultant beats an agency on custom software for small business
When you need judgement rather than capacity: someone senior to assess, recommend and leave your team able to act. A consultant also carries a political function, since an outsider can say things an employee cannot, and that is a legitimate reason to hire one. If what you actually need is people to build something, that is capacity and an agency or contractors will cost less and deliver more.
The deliverable from custom development companies should be usable by the people who stay
A recommendation your team cannot act on is a failed engagement regardless of its quality. Require the output to name specific next steps with rough effort, to say what was considered and rejected and why, and to be written for the people who will do the work rather than for the executive who signed the invoice. Ask to see a redacted example from previous work before committing.
Keep the engagement honest about its own limits
A consultant seeing your systems for two weeks knows less about them than your team does, and the value is perspective rather than depth. Good consultants say so and structure their work around it, interviewing widely and testing conclusions with the people who will live with them. Be wary of confident recommendations delivered without having spoken to the engineers who maintain the thing being recommended about.
Questions people ask about software consulting
When should we hire a software consultant?
When you have a decision to make and lack the information or experience to make it, and can write the question down. If what you need is people to build something, that is capacity and an agency or contractors will cost less.
How do we stop consulting ending in an unused document?
Scope it around a written question with an agreed shape of answer, and require the output to name specific next steps with rough effort, written for the people who will do the work rather than for whoever signed the invoice.
How much can a web consultant really know in a short engagement?
Less about your systems than your team does. The value is perspective, not depth. Good consultants say so, interview widely and test conclusions with the engineers who will live with them.