Every small business owner I talk to has the same blind spot when a vendor sends over a proposal. They try to evaluate the technology. Is React better than WordPress? Is this hosting setup actually secure? Does "AI-powered" mean anything here, or is it just this year's word for "we added an if-statement"?
You can't evaluate that, and you don't need to. You can evaluate the thing you already evaluate every time you buy a truck, hire a manager, or sign a lease: whether the money makes sense. You do this competently in every other part of your business. The tech industry has just convinced you it's different because the vocabulary is unfamiliar.
It isn't different. A proposal is a capital spending request wearing a hoodie. Treat it that way.
How to Evaluate a Technology Proposal: Four Numbers
Technology ROI for a small business isn't a special discipline. It's four numbers. Before you sign anything, ask the vendor for them in writing, in the proposal or in an email you can point back to later.
- Total first-year cost, including the recurring pieces. The headline number on most quotes is the build cost. It rarely includes hosting, the monthly plugin subscription, the SEO retainer they mention in passing on page four, or the "optional but recommended" maintenance plan that turns out to be not very optional once something breaks.
- Then ask what it costs in years two and three, because a lot of proposals are priced to look cheap in year one and expensive forever after. Get both numbers before you commit to either.
- What it replaces, in dollars or hours. If this system is supposed to save your office manager six hours a week, say so and put a number on those hours. If it's supposed to replace a $400/month tool you're already paying for, name the tool.
- Payback period, in months.
That last one is the number vendors skip most often. If a vendor can't produce all four, it usually means nobody ran the math before writing the proposal. You're being asked to trust the pitch.
What Payback Period Actually Means
If you've never used this term, it's simpler than it sounds. Payback period is just: how long until this thing has paid for itself?
Say a landscaping company in Gilbert gets a new website for $6,000. The vendor claims it'll bring in leads. Vague, so you push for a number, and you land on two extra jobs a month at $400 average margin each. That's $800 a month. But the site isn't free to keep running. Say hosting and maintenance run $100 a month, which is a common agency retainer for the two together, though plenty of shops (mine included) charge a fraction of that. Use whatever your vendor actually charges. At $100, the money coming back to you is $700. Divide $6,000 by $700 and you get about 8.6 months. After that, the site is just making you money.
That's a payback period. It's one division problem, and the only real work is remembering to subtract the recurring costs before you divide. It ignores the time value of money and everything that happens after the payback date, which on a decision this size doesn't matter much. Anyone who's run a business can do it in their head.
So when you're trying to figure out whether a website quote is reasonable, work the payback period before you react to the sticker price. A $15,000 booking system that retires a $900-a-month software subscription and six hours a week of front-desk admin pays for itself inside a year, and that's after subtracting what the new system costs to run. A $4,000 site that brings in nothing never does.
Sometimes the math genuinely works, and payback inside a year on a build like that is a good outcome. The red flag is when nobody can tell you where the number came from. If the vendor can't say what the two extra jobs a month are based on, or waves at "increased visibility" instead of a figure, you've learned something important about how carefully this was thought through.
The Question That Does the Most Work
Here's the single most useful thing you can ask a vendor, tech or otherwise: "What would have to be true for this to be a bad decision?"
A vendor who's actually run the numbers has an answer ready. Maybe it's "if your close rate on leads drops below 15%, this doesn't pay back inside a year." Maybe it's "if you don't have someone updating the blog monthly, the SEO piece won't do much." Good vendors know where their own pitch is fragile, because they've stress-tested it.
A vendor who says "nothing, really, this is a great investment for any business" has told you they haven't thought about it critically, or they're hoping you won't either. Neither is a good sign.
If the vendor in question is a marketing agency, the money questions are the same but the reporting games are different. I wrote up five questions to ask an agency before you sign that sit on top of these four numbers.
"It's an Investment in Your Brand" Should End the Meeting
Every business spends on things it can't fully quantify. Brand is real. So is goodwill, and so is the vague sense of trust a polished site creates before a customer ever calls.
But when a vendor uses "investment in your brand" to explain away the absence of those four numbers, it's what gets said when there isn't a real answer. A vendor who genuinely believes in the brand value of their work can still tell you what it replaces or what it's supposed to move. If the response to "what's the payback here" is a sentence about brand instead of a number, write that down. That's the answer.
Where the Payback Math Breaks Down, Honestly
Not everything should be evaluated this way, and pretending otherwise causes real damage. Backups and their offsite retention, security patching, email deliverability infrastructure, uptime monitoring, domain renewal: these aren't investments with a payback period. They're insurance. The ROI on insurance rarely pencils out on paper. You buy it anyway because the downside of skipping it is a business-ending event, and the odds of that event aren't zero.
Risk is the test for insurance spending. Ask what happens to your business if this fails and you didn't have it. A ransomware attack with no backups, a domain that lapses because nobody was watching: these cost businesses far more than the "insurance" would have.
The failure mode I actually see is owners applying payback-period thinking to insurance-shaped spending, concluding it doesn't pencil out, and skipping it. That's backwards, and it's how a backup plan you declined at thirty dollars a month becomes a recovery quote you have to think hard about and a week you spend not selling anything.
My Own Version
I came to this from finance. My master's is from Vanderbilt's Owen Graduate School of Management, earned long before I was building websites and systems for small businesses, and it's why a proposal from me arrives with the payback math already in it. What it costs this year, what it costs after, what it's supposed to replace, and how long before it pays for itself. If the numbers don't clear, I'll tell you that instead of finding a reason they should.
Do this today: pull up the last tech proposal you signed, or one that's sitting in your inbox right now, and try to answer the four numbers yourself. If you can't, that's not a personal failing so much as a sign the proposal was never built to be evaluated in the first place.
Either way, send it over. One proposal, mine or anyone else's, and I'll run these four numbers on it and get them back to you, usually within a couple of business days. Start on the contact form and I'll ask you for the file, or email it straight to hello@marshland.software. No charge, no pitch attached. If the proposal holds up, you'll know why. If it doesn't, you'll know before you sign.
Related reading: You Get What You Pay For in Web Design, Custom Software for Service Brands, Why Web Hosting Is So Expensive (And What To Do About It), and The Real Cost of 'Free' Website Hosting.