Build, buy or leave alone: a decision framework for a ten-person firm

The decision goes wrong when it starts from a product demo or a technology. Start instead from an inventory of what the firm actually does repeatedly: origination, buyer research, outreach, document production, pipeline reporting, data room management, engagement administration.

For each one, estimate three things honestly: how many senior hours it consumes in a typical month, what an error in it would cost, and how often it sits on the critical path of a live deal. Those three numbers rank the workflows worth touching. A task that is manual but rare, cheap and safe is not a problem, whatever a vendor says about it.

Some workflows should not be modernised. If a process consumes few hours, produces few errors and the person running it is not complaining, the cost of change, in training, migration and disruption, exceeds any saving. A spreadsheet that one person maintains well, with nothing depending on it downstream, is not a liability. It is a solved problem.

The discipline here is refusing improvement for its own sake. Every system added to a small firm carries a permanent attention cost, and that budget is better spent where the pain is measurable.

Accounting, payroll, e-signatures, scheduling, video calls, standard virtual data rooms: your firm gains nothing from doing these differently from everyone else. Vendors spread the engineering cost across thousands of customers, the products are mature, and the switching costs are low.

The test is simple: if your firm does this task the same way every other firm does, buy the standard tool and move on. Bespoke software in a generic workflow is expense without advantage.

A small number of workflows are the firm: how you originate, how you research and approach buyers, how you assemble documents, how partners see the state of every mandate. This is where off-the-shelf products force a template, and where the gap between the template and how your firm actually works gets absorbed, silently, by senior time.

Building is also the defensible choice when confidentiality constrains where data can sit, so pooled multi-tenant software is unacceptable, and when the real cost is not any single tool but the manual work in the gaps between several tools. Integration problems rarely have an off-the-shelf answer, because your combination of tools and habits is specific to you.

Buying costs more than the subscription. Add the hours spent working around the template, the exports and re-imports, and the administrative effort of forcing your process into someone else's model. That is the number to compare, not the licence fee.

Building costs more than the build. Software that is not maintained decays into exactly the failure we describe elsewhere on this site: quietly abandoned within a year. If your firm is not prepared to pay for ongoing maintenance, do not build. Buy the imperfect tool instead, because an unmaintained bespoke system is the worst of both options. That is a legitimate answer, and for most workflows the right one.

The framework is easiest to see applied to the workflows most small deal firms have in common. The answers below are the ones we reach most often; a particular firm's numbers can move any of them.

Notice how short the build column is. Most of a firm's processes end up bought or left alone, which is what makes the remainder worth building properly.

Two things, and the second matters more than the first. The first is an initial build: scope and price agreed up front, working software within weeks, used on live mandates while it is refined. The second is maintenance. A bespoke system that nobody keeps current decays into a spreadsheet within a year, so building is only the right answer for a firm prepared to pay for a steady cadence of small changes afterwards.

Verdalyze structures that as a retainer billed every 4 weeks on rolling terms, with no per-seat surcharges and no lock-in. The structure is less important than the principle: whoever builds the system should expect to still be adjusting it a year later, and a firm that is not prepared for that should buy the imperfect tool instead.

Once a quarter, not once. Workflows move between columns as the firm grows: the spreadsheet that was fine at six people becomes a liability at twelve, and the standard tool that fitted last year forces a template this year. The inventory, the three estimates and the four questions take an afternoon to revisit, and the output is a short list of what to change and a longer list of what to leave alone.

Ask whoever advises you to show their working. Every recommendation to build should arrive with the buy and leave-alone alternatives spelled out and the reasoning written down for the firm to challenge. Building is what we sell, which is exactly why we only recommend it when it clearly beats the alternatives, and why the reasoning has to be on the table.

Four questions settle most cases. Does this workflow consume meaningful senior time every week? Does doing it our way help win or protect mandates? Does confidentiality constrain where the data can sit? Are we prepared to maintain what we build?

Four yeses point to building. Fewer than four usually point to buying, or to leaving the workflow alone. Applied honestly, the test rules out bespoke software for most of a firm's processes, which is exactly why it is worth trusting where it rules it in.