Say you run a two-person landscaping company. Every month you lose about five hours to writing quotes and chasing follow-ups, and an AI tool offers to take that off your plate for $50 a month.

The math looks simple. Five hours, fifty dollars. Each reclaimed hour needs to be worth more than $10 for the tool to beat what you're paying for it.

That's the right place to start and the wrong place to stop. It doesn't count the afternoon you spend setting the thing up, the time you spend checking what it writes, or the second seat you add in month two. Count those and the threshold climbs — often past $18 an hour.

Software subscriptions add up quietly. A $50 fee disappears next to a $4,000 month. Five of them don't. For small businesses, every recurring tool should earn its place.

This guide covers what belongs in total cost, how to price an hour of your own time honestly, how to run the calculation, what your P&L can and can't confirm, and how to test the whole thing over 90 days before the renewal date.*

What is software ROI, and how do you measure it?

Return on investment measures the net benefit an investment produces relative to its total cost. Software ROI applies that measure to a tool you pay for every month — which is a fairer test than it sounds, because a subscription keeps charging you whether you open it or not.

ROI (%) = [(Total benefits − Total costs) ÷ Total costs] × 100

Three outcomes:

1. Positive ROI. Total benefits exceed total costs. The tool earns its place and delivers a positive return.
2. Zero ROI. Benefits equal costs. The tool breaks even — no better off, no worse.
3. Negative ROI. Costs exceed benefits. The tool costs more than it returns.

Two more financial metrics show up in the same conversation, and people use them interchangeably. They aren't.

Break-even is the point where benefits and costs meet. Here it's expressed as a break-even hourly rate: what one reclaimed hour has to be worth before the tool pays for itself.

Payback period is how long it takes to get there — how many months of net benefit it takes to recover your initial investment.

ROI is neither. It's a percentage: how much you got back relative to what you put in.

An accurate ROI calculation depends far more on realistic inputs and actual performance than on a complicated formula. The rest of this guide is about the inputs.

What costs belong in a software ROI calculation?

The subscription is the smallest part of what a new software system costs you. Total costs also include implementation costs, the labour you spend learning and reviewing, ongoing operational costs, and what it takes to leave.

• Subscription. $50/month, or $600/year, before taxes, usage charges, or price changes.
• Setup and implementation costs. Account setup, configuration, data migration, integrations, writing prompts or templates, and workflow redesign.
• Learning and oversight labour. Reading documentation, testing output, training whoever else touches it, writing down the new process, approvals, and troubleshooting.
• Ongoing operational costs. Additional seats, usage overages, paid add-ons, maintenance, security checks, and human review time.
• Duplicated subscriptions. A tool that overlaps something you already pay for is an additional cost with no matching benefit — and the easiest one to miss, because the duplicate usually sits inside another tool's feature list rather than under its own line item.
• Switching and exit costs. Exporting data, replacing workflows, retraining users, losing access to stored work.

One-time costs don't disappear because they happened once. Spread them across the months you're measuring — accountants call it amortizing. The landscaper's four hours of setup at $30/hr is $120, and across a three-month pilot that's $40/month on top of the subscription.

Monthly total cost = subscription + monthly usage and add-ons + amortized setup and training + ongoing review and maintenance

Pick that evaluation period deliberately. It moves your monthly total more than almost anything else here. Spread the same $120 over twelve months and it drops to $10/month — the same tool looks roughly twice as good. Three months is the honest window while you're still deciding. Once the pilot ends and the tool stays, the setup cost is behind you.

Does a low price mean good ROI?

No. A low list price doesn't guarantee a good ROI. A new system that creates extra work, sits unused, or duplicates something you already have can produce negative ROI while looking cheap. Most owners find that out at renewal, when the charge shows up and the tool hasn't been opened since March.

How much is five hours of your time actually worth?

To put a realistic dollar value on hours saved, multiply your verified net hours by a conservative hourly value — an employee's loaded labour cost, a contractor rate, or what the time is worth if it reliably goes to billable work.

Five hours saved is a claim, not a result, until you verify it.

Establish a baseline first. Identify the exact task — for the landscaper, that's writing quotes — how often it occurs, and how long it currently takes without the software. Time it. Don't estimate it.

Then time the same job again after adoption — counting everything: prompting, reviewing output, correcting errors, transferring data, fixing whatever came back wrong. What's left is your net time saved.

Monthly value of time saved = verified net hours saved × conservative hourly value of that time

Where review time goes: count it once. If you measure hours net of reviewing and correcting, don't also add review time to your cost side. If you'd rather measure gross hours, charge review time as an operational cost. Either works. Doing both quietly halves your result.

Four ways to price the hour:

1. Employee loaded labour cost: What it costs to employ someone for an hour, including benefits and overhead.
2. Contractor rate: What you'd pay someone else to do the same task.
3. Owner replacement cost: What you'd pay to hire someone to do it if you weren't doing it yourself.
4. Opportunity value: What the hour is worth if it reliably goes to billable or revenue-producing work.

Be conservative about whether you'll actually use the tool, and how much. If you're not certain the reclaimed time gets reallocated to higher-value work, don't value it at your top billable rate.

Expect results to vary more by person than by tool. In Generative AI at Work — Brynjolfsson, Li and Raymond, published by the Stanford Digital Economy Lab — researchers studied 5,179 customer support agents using an AI assistant. Average productivity rose 14%, but nearly all of it went to the least experienced workers, who gained around 34%. Experienced staff gained little. One study, one task, one workplace: evidence that gains are measurable and uneven, not a forecast for your business.

Does time saved count as profit?

No. Time saved is capacity, not net profit or automatic cash savings. The hour only produces financial returns if it reduces a cost, prevents a hire, increases output, improves service, or gets reallocated to revenue-producing work. An hour absorbed by low-value tasks doesn't change your numbers.

How do you calculate software ROI for a $50 AI tool?

To calculate software ROI on a $50/month AI tool, multiply your verified net hours by a conservative hourly rate to get total benefits, subtract your total costs, divide by total costs, and multiply by 100.

Back to the landscaping company: five verified hours a month, an hour valued at $30, four hours of setup. Calculate ROI in two passes — the simple version, then the one that includes everything else.

Pass one — subscription only:

1. Verified net hours saved: 5/month.
2. Hourly value: $30.
3. Monthly benefit: 5 × $30 = $150.
4. Net benefit: $150 − $50 = $100.
5. Apply the ROI formula: $100 ÷ $50 × 100 = 200% simple monthly ROI, before additional costs.

THE SIMPLE BREAK-EVEN
$50 ÷ 5 hours = $10 per hour A simplified example, before hidden costs. If an hour of your time is worth less than $10, the tool loses money on the subscription alone.

At a $10 hourly value, five hours returns $50 against a $50 subscription: exactly break-even.

Pass two — total cost:

Same tool, same price, same five hours. The threshold nearly doubled once implementation costs were counted.

That last row is the one to check yourself against. Nothing changed except the hours the landscaper actually verified, and the tool went from earning its place to costing $30 a month.

A higher ROI percentage can be mathematically correct when the number you divide by is small. A $5 tool returning $15 shows 200% and fifteen dollars.

Examine adoption, output quality, risk, and the absolute financial impact before you renew. Percentages alone don't make informed decisions.

Can your P&L show whether a software investment is working?

Partly. Your P&L confirms what you spent. It can't tell you how much profit the tool actually added.

Software subscriptions generally appear as operating expenses — often coded to "Dues and subscriptions," depending on your chart of accounts and your accountant's guidance. Once categorized, comparing them across periods is straightforward. Pull your profit and loss statement for three consecutive months, or two quarters, and watch the effect on total operating expenses and net profit.

What your P&L can verify:

• Recorded software costs, and when you paid them.
• Revenue trends across comparable periods.
• Movement in total operating expenses and other expenses.
• Changes in profitability over the period you select.

What can't your P&L prove?

It can't prove which tool caused an increase in total revenue, whether five hours were truly saved, or whether output quality and customer satisfaction improved. Correlation isn't causation — seasonality, a price change, or a staffing change moves the same numbers just as much. A good spring moves them more than any tool will.

So pair it with other metrics. Keep a short list of what you're tracking — an operating scorecard: hours saved, output volume, error and rework rate, response time, how often you open the tool, and any revenue or savings you can fairly trace to it. That list proves what the P&L can only hint at.

It takes both records. The financial side and the operational side answer different halves of the same question, and neither is worth much alone. Which is the argument for keeping them in one place rather than rebuilding them from three every time you need an answer. If you're not tracking this cleanly yet, here's how to track business expenses, and how to understand your P&L in Wave.

PRO-TIP FOR WAVE USERS
Toggle your P&L to Accrual Basis when evaluating a tool's ROI. This matches your monthly software expenses against the revenue earned during that same period, even if your clients haven't settled their invoices yet.

One thing neither record shows: timing. A tool can return positive ROI over a year and still land badly in a slow month. Run the cash flow formulas before committing to a recurring expense you can't comfortably fund.

How do you count revenue and intangible benefits without double-counting?

Categorize your returns before you total them. Combining different types of return without tracking their source is how ROI calculations get inflated. Returns fall into five categories:

1. Direct cost savings: Fewer contractor hours, less overtime, a subscription you can finally cancel.
2. Increased capacity: More work finished in the same hours — the capacity you'd otherwise have to hire for, which is one honest way to increase efficiency.
3. Increased revenue: New sales or faster delivery you can trace to the tool and not to a busy spring.
4. Avoided costs: Expenses you didn't incur because the tool caught something first.
5. Intangibles: Leads answered the same day, fewer errors to walk back, quotes that read the same every time, customers who don't have to chase you.

That last group tends to show up as customer satisfaction long before it shows up in revenue, which is exactly why it's hard to price.

Count customer retention or expected revenue only when there's a defensible link to the specific software and enough data to estimate the effect. "We grew, and we also bought this" is not a link.

Count the reclaimed time either at its hourly value or as the revenue it generated — never both. Valuing five hours at $30 and then adding the revenue those same five hours produced counts one outcome twice. Pick whichever you can measure more defensibly.

Track intangible benefits on a separate scorecard rather than assigning arbitrary dollar values. Faster customer interactions are worth tracking. They don't belong in the ROI calculation itself unless you can measure what they earned or saved.

Productivity isn't the only thing worth reviewing. Before widening any tool's role, check where your data lives, who owns it, whether it trains someone else's models, and how you'd get it back out. NIST's voluntary AI Risk Management Framework covers the same ground in more detail — reliability, security, privacy, accountability — if you want the long version. The same questions apply anywhere generative AI can help your business.

When the heavier methods earn their keep

Not here. A $50 subscription doesn't need them. A higher investment does — custom software, larger projects, replacing a system the business runs on.

On larger software projects, development costs and project management time run across a whole software development process, and the money is tied up for years. That's where net present value and discounted cash flows earn their complexity: they weigh the time value of money against long-term business objectives. On a subscription you can cancel on Tuesday, they don't.

How long should you run a software pilot before measuring actual ROI?

Measuring ROI over a 30-, 60-, and 90-day pilot moves you from projected ROI to actual ROI: verify adoption at 30 days, compare hours saved against your assumptions at 60, and calculate measured ROI and payback period at 90.

Projected ROI tells you what should happen. Measured ROI tells you what did. The gap between them is where most subscription decisions go wrong.

Before you start, pick one job you repeat — the landscaper's quotes will do. Record how long it takes now and how good the output is, set a ceiling on total cost, and write down the benefit you expect in numbers.

At 30 days, you're checking one thing: are you actually using it? Record setup effort, output quality, and early time savings. If the tool is sitting untouched or producing errors you keep fixing, repair the job before you expand it. Expanding a broken process just multiplies it.

At 60 days, the question is whether your assumptions held. Compare actual hours saved and additional costs against what you guessed at the start. Look for rework, tools that don't talk to each other, and jobs you're now doing twice. Then recalculate total cost from what you've actually spent, not the sticker price.

At 90 days, you have enough to decide. Calculate actual ROI and payback from measured benefits and confirmed total costs, then put the relevant P&L periods and your operating scorecard side by side.

Then pick one action: keep the tool, expand it, renegotiate or downgrade, change the job and retest, or cancel.

Measuring ROI is an ongoing process, not a one-time calculation. Usage, pricing, team size and revenue all move. A scheduled review before each renewal turns a guess into a decision you can back with data — and it's the cheapest way to raise the return on everything you already pay for.

Should you keep, improve, or cancel the subscription?

Run this before you renew any software investment. Six checks, and the tool has to clear all six.

1. It solves a job you actually do every month — not one you might do.
2. You've verified actual usage — hours saved, counted after the time you spend reviewing and fixing.
3. Total cost includes everything — subscription, add-ons, training, integrations, maintenance, oversight.
4. The saved time went somewhere — lower costs, higher-value work, better service, capacity you'd have hired for, or more revenue.
5. You know what happens to your data — who can see it, who owns it, how you'd get it back, and who checks the tool's work.
6. No negative-ROI signals — low adoption, overlapping tools, new manual work, unreliable output, rising usage costs, or no measurable improvement.

Then take one clear action: keep and measure, improve the workflow and retest, downgrade, replace, or cancel.

Keeping an unused subscription because cancelling feels like an admission of failure is not a financial strategy. If the tool isn't producing a measurable result, the cost is certain and the benefit isn't.

If you're weighing several software investments at once, the same discipline applies when you systematize and automate more of how the business runs.

Conclusion

Learning how to calculate software ROI starts with total costs, verified benefits, and conservative assumptions — not the subscription line and a vendor's promise.

Back to the landscaper. A $50 monthly tool saving five hours needs each reclaimed hour to be worth more than $10 before hidden costs to move beyond break-even. Count the four hours of setup and that threshold climbs to $18. Verify the hours instead of accepting them and it moves again.

Your P&L validates recorded cost and overall financial direction. Your operational metrics prove whether the tool created the expected result. Neither works well alone, which is the case for keeping them in one place rather than three.

So run a measured pilot. Compare your P&L periods. Calculate actual ROI before the next renewal — rather than relying on a sales claim or keeping unused subscriptions indefinitely.

Wave's accounting software keeps your expenses categorized and your P&L current, so comparing three months is a report you pull rather than an evening you lose.