What CFOs Should Look for in Software ROI

Most software ROI calculations focus on the wrong number. The license fee is the visible cost, so it’s the one that gets scrutinized. But the license fee is almost never the largest cost of running a business application. The hidden numbers — productivity loss, error rates, workaround maintenance — usually dwarf the line item on the contract.

If you’re a CFO trying to evaluate a software investment, here’s the math worth running.

The four-input formula

ROI on business software comes down to four inputs:

• License or build cost (the easy number)
• Productivity gain: time recovered per user per week × loaded labor cost
• Error reduction: errors per month × cost per error
• Workaround maintenance cost: time spent on spreadsheets and shadow systems the software was supposed to replace

The first one is on the invoice. The other three live in your operating expense and never show up under “software.” That’s the problem. The conversation about software cost happens against the smallest of the four numbers.

Why the productivity number gets underestimated

Software vendors love productivity claims, so CFOs reflexively discount them. That’s healthy. But it leads to systematically underestimating the real productivity gain when the software fits.

The mistake is modeling productivity as “current admin cost minus future admin cost.” This is incomplete. The bigger productivity number is what your team starts doing with the time the new tool gives back. A salesperson with two extra hours a week doesn’t fill them with admin. They sell more. A finance team with their month-end close cut from five days to two doesn’t take three days off. They start producing analysis instead of reports.

The honest way to model it: estimate the time recovered, then multiply by the marginal value of the role, not the average labor cost. Sales reps get the deal-margin multiplier. Finance gets the analysis-output multiplier. The numbers usually surprise the CFO who runs them for the first time.

Error cost is where the big numbers hide

A 0.5% error rate sounds small. Multiply across a hundred quotes a month at $500 cost per error and that’s $30,000 a year for one team in one workflow. Multiply across the business and the number is usually in the six figures.

Custom software isn’t always cheaper on license. It’s often cheaper on error cost. The system is designed for your specific error modes. The exceptions you keep hitting in the generic tool — the ones that require manual override every time — are first-class citizens in custom software.

The workaround maintenance line item

Every off-the-shelf tool that doesn’t quite fit gets a spreadsheet built around it. That spreadsheet is software. It has a maintenance cost. The person who built it spends time keeping it current. New hires need to be taught it. It breaks when one of the underlying systems changes. When that person leaves, their successor inherits a system nobody documented.

This is hidden software cost. It’s not on the SaaS budget. It’s in salary lines for people doing systems work that wasn’t their job description.

The three-year frame

ROI calculations get distorted by one-year thinking. License fees are annual. Build cost is one-time. The right comparison is over a three-year horizon, because that’s the realistic life of the software.

A $35,000 one-time custom build vs. $30,000 a year in SaaS fees. Year one, the custom build looks more expensive. Year two and three, the SaaS bill keeps coming. By month 14 the custom build is cheaper. By month 36 it’s saved you $55,000 and you still own it. (See Why Custom Software Is No Longer Just for Enterprise for what changed about the build option.)

What this means for your next software decision

When the next software approval lands on your desk, ask three things:

• What’s the total annual cost — license, plus productivity loss from gaps, plus error cost, plus the workaround maintenance line nobody tracks?
• Over three years, what does that compare to in a custom build that fits exactly?
• If we don’t change anything, what does year three look like?

The third question is the one most CFOs don’t ask. It’s also the most important. Inaction has a cost. The numbers usually argue for changing something.