The 80% Rule for Buy Versus Build Software | Active Logic Insights

Every buy-versus-build conversation I have with a mid-market CEO ends at the same point: “we know our current software isn’t right, but custom feels like too much.” The decision has a clean threshold. If off-the-shelf covers 80% or more of what your business actually needs, use it. Below 80%, the math flips, and most companies do not realize it has already flipped on them.

The Rule, Stated Plainly

If a standard off-the-shelf product covers at least 80% of your real workflow and you can afford the license fees, do not build custom. Custom software is heavier, slower to deliver, and a longer commitment than buying a SaaS seat. When the off-the-shelf product actually fits, picking custom over it is vanity, not strategy.

Below 80%, the rule inverts. The cost of forcing a product that fits 60% or 70% of your needs to look like a product that fits 95% of them is usually two to three times what most leaders estimate, and almost none of that cost goes through the line item labeled “software.”

I have given this rule to roughly every prospect I’ve spoken to in the last decade. It is the cleanest single threshold I know for the decision, because it forces the conversation onto what the buyer can actually measure: features covered, workarounds running, license fees paid. Gut feel about whether your business is “unique enough” to deserve custom is the wrong frame, and it produces wrong answers in both directions.

Why 80%, Not 50% or 95%

The threshold lives at 80% because that is the point where the hidden costs of forcing off-the-shelf to fit become structurally larger than the cost of building custom. Above 80%, the gaps are small enough that an engineer-hour spent customizing the off-the-shelf product produces real leverage. Below 80%, the gaps are large enough that customizations stack, integrations multiply, and shadow workflows compound until the total annual cost matches or exceeds a one-time custom build.

A 95% threshold sounds tempting (why not insist on near-perfect fit?) but it sets a bar that almost no real software meets, including custom. Every shipped product has gaps. Choosing custom because off-the-shelf is at 88% rather than 95% gets you a system that is at 92% when delivered, not 100%, and you paid six figures for the upgrade. Off-the-shelf at 88% is the better trade. The judgment matters at the lower end of the range, not the top.

The Three Places You Pay Below 80%

Off-the-shelf below the threshold costs you in three line items at once, and the combined total is what makes custom cheaper than the math suggests.

First, you pay the license fee on software you cannot fully use. SaaS pricing is per-seat or per-feature-tier. You are paying for the whole product even when your workflow uses 60% of the surface area. That fee compounds every year with annual escalators that most companies forget to negotiate against.

Second, you pay customization fees to bend the product toward your business. Sometimes this is a configuration services contract. Sometimes it is a Salesforce engineer or Dynamics consultant on your payroll. Sometimes it is a dedicated implementation partner billed at $200 an hour. The cost looks like services, not software, which is why it never lands in the same conversation as the license fee.

Third, you pay for shadow infrastructure: the Zapier integrations, the Excel sheets passed around the operations team, the manual data entry between systems that don’t talk to each other. This cost shows up in headcount and in error rate. The bookkeeper who spends ten hours a week reconciling two systems is the cost of a software gap, even though that hour gets charged to the operations budget.

Add those three together, project them across the five-year life of your software, and the number is almost always larger than what a custom build would have run. The trap is that no single line item looks expensive in isolation.

The Salesforce Math

Salesforce is the canonical example of how the three-cost stack plays out. Mid-market companies sign Salesforce because their board, their lead investor, or their head of sales has been told that serious companies run Salesforce. License fees run six figures per year by the time the seat count climbs. Five-to-ten-year contracts are the norm.

Then the company discovers the platform does not quite fit its pipeline. So it hires Salesforce engineers, internal or contracted, to customize the product. We have clients in insurance and pet insurance who have been paying Salesforce engineering salaries that, when totaled, match or exceed what a custom CRM build would have cost outright. They’ve spent years and millions of dollars customizing a product they do not own, cannot take with them, and cannot meaningfully extract their data from without another seven-figure migration project.

The migration cost is the lock-in. By the time leadership realizes the platform does not work for them, they have built so much business logic on top of it that getting out costs more than starting over fresh. They are stuck paying license fees, paying customization fees, and paying for the shadow Zapier-and-Excel layer that fills the remaining gaps, on a product they will never own. That is exactly the three-cost stack the 80% rule predicts.

The pattern repeats across Microsoft Dynamics, NetSuite, HubSpot Enterprise, and every other “configurable” enterprise platform. The product is configurable until the configuration costs exceed the cost of just building your own software, and most buyers do not recognize that crossover point until they are years past it.

The Real Test Is Process, Not Industry

Buyers ask me whether custom is right for their industry, and the answer is almost always that industry is the wrong variable. Process is.

Two landscaping companies illustrate this. A company that runs a standard mow-and-bill operation on a single-market basis fits off-the-shelf scheduling and invoicing software cleanly. Most operators in their industry do roughly the same thing the same way, and the SaaS market has produced products that cover 85% to 90% of the workflow at a reasonable price. Custom software is the wrong call for them.

The other landscaping company we worked with runs a franchisee model with non-standard service tiers, regional pricing variability, and operational structures the standard products did not anticipate. The off-the-shelf options covered maybe 50% of their actual workflow. They had been running on a combination of three SaaS tools, an internal Excel layer, and a Zapier pipeline that broke quietly every quarter. We built them a custom operations platform that consolidated the three tools and retired the Excel layer.

Same industry. Different process. Different answer. The 80% rule does not care about the industry label. It cares about whether your specific workflow looks like what most companies in your market do, or whether your business has structural differences that off-the-shelf cannot accommodate.

The Lost-Revenue Test the Rule Misses

The 80% rule, applied naively, counts features. Feature coverage is the floor of the analysis, not the ceiling. The harder question for established companies: does the off-the-shelf product enforce your business process, or does it merely accommodate it?

A B2B IT and AV company we worked with, around 30 to 50 employees, was running off-the-shelf proposal software that technically covered all the features they needed. Salespeople could quote anything. But the software did not enforce completeness, and on hardware-heavy IT projects, salespeople were forgetting line items routinely. Rack hardware. Cable management. Edge switches. Contracts got signed and then the company discovered it had under-quoted by thousands of dollars on each project. The off-the-shelf product was at maybe 85% feature coverage, which would have passed a naive 80% audit. The actual workflow gap was much wider.

We built them a custom estimating tool that combined a trained model with LLM-driven analysis of their historical estimates. When a salesperson started a new quote, the software compared it against similar past jobs and surfaced what was likely missing. The company stopped leaving money on the table because the software encoded institutional knowledge that had been sitting in employees’ heads.

When running the 80% audit, count both features (what can the software do?) and process enforcement (does the software make my team get it right?). A product at 85% feature coverage but 50% process enforcement is below the 80% threshold once you do the harder count.

When the Rule Says Stay Off-the-Shelf

Two cases where the 80% rule almost always lands above the threshold, and custom is the wrong answer.

The first is accounting software. State and local tax compliance complexity is too high, regulatory change is constant, and the ROI on custom accounting almost never exists outside the Fortune 100. Use QuickBooks, Xero, NetSuite, or a regulated specialty product. We use QuickBooks ourselves. The hard no on custom accounting is one of the cleanest decisions in the buy-versus-build space.

The second is truly commoditized operations. Businesses where most operators in the industry run the same workflow the same way (lawn care, basic contracting, residential remodeling, dental practice management) sit firmly above 80% on standard SaaS. The exception is when the company has a non-standard model layered onto the commoditized base, in which case the rule trips back below 80% and custom becomes the answer. The default is off-the-shelf, and pushing past it without a clear non-standard process is buying problems you do not need.

How to Run the Audit

Four steps. Do this with your operations leader and your finance leader in the room together.

• List the workflow. Write down the actual day-to-day operational sequence your team runs. Not the diagram in your strategy deck. The thing that happens when a real customer order, a real project, or a real service request enters your business.
• Mark the gaps. For each step, mark whether your current software handles it, partially handles it, or doesn’t handle it. Be honest about the partial-handle category, because that is where the shadow infrastructure hides.
• Cost the workarounds. For each partial-handle and doesn’t-handle step, estimate the hours per week your team spends bridging the gap. Multiply by loaded labor cost. Project across five years.
• Compare totals. Add the five-year license cost, the five-year customization cost, and the five-year workaround cost. Compare that total to a one-time custom build amortized over five years. The numbers are usually closer than buyers expect.

The audit takes a half-day if done well. The output is the actual buy-versus-build math for your specific business, not the generic version, and it tells you which side of 80% you are sitting on.

The 1-to-2 Year ROI Test

The final filter on the 80% rule: can your business afford a custom build, and will it pay for itself in roughly one to two years through efficiency gains, revenue capture, or competitive advantage? If yes, the rule favors custom and you should move. If no, off-the-shelf is the right answer until your revenue catches up to support the investment.

This is the test that pulls early-stage and sub-$10M-revenue companies off custom. Below a certain revenue threshold, the math does not work even when the 80% audit says custom should win. The right call there is to run on off-the-shelf, accept the workarounds, and revisit the audit annually as the business grows. Custom software paced to a partner Team as a Service engagement starts making sense once the workaround cost crosses the threshold the business can fund.

What to Do Next

If the audit lands clearly above 80%, stay on off-the-shelf, negotiate aggressively on license escalators, and move on with your day. If it lands clearly below 80%, the next questions are what custom actually costs (the cost article covers honest ranges from freelancer through enterprise) and how the build actually runs (the process article walks through every phase from first call to launch). If you are sitting on the threshold, the deepest treatment of the broader trade-off is in the pillar, what is custom software.

Most companies discover, when they run the audit honestly, that they crossed below 80% about two years before they noticed. The cost of the delay shows up in retained workarounds, in retention, in lost revenue from process enforcement gaps, and in the competitive deficit against operators who built custom earlier. The audit costs a half-day. The procrastination is the expensive part.