What Counts as a Conversion? Deciding Your Website's One Job

Ask a room of small business owners what their website is for and you’ll hear the same word doing a lot of heavy lifting: presence. The site exists so the business exists online. I get it, and there’s some truth in it. But a website that only exists is a parked car. It cost you money and it isn’t taking you anywhere.

I wrote a longer guide on this, what a conversion-optimized website actually is, and the first question in that guide deserves its own post: what actually counts as a conversion for your business? Until you can answer that in one sentence, nothing else about your website can be judged. You can’t tell whether the design is working, or the copy, or whatever someone’s charging you monthly to “manage” it. There’s no such thing as “working” until the site has a job.

So let’s figure out the job.

A conversion is just “they did the thing”

Strip away the jargon and a conversion is a small, plain event. A visitor arrived on your site, and then they did the thing you wanted them to do. That’s the whole definition. The entire discipline of conversion optimization hangs off it.

The word comes from marketing, where a browser “converts” into a lead or a customer, the way water converts into ice. Same person, new state. Before the visit they didn’t know you or hadn’t acted. After the visit they’ve called you, booked with you, bought from you, or put your address into their maps app. Something changed, and your website caused the change.

Notice what the definition quietly requires, though. There has to be a thing you wanted them to do. If you never chose one, conversions can’t happen on your site, in the same way you can’t score in a game with no goal posts. Visitors still show up. They skim a bit, they leave, and whether that visit was a success is anyone’s guess. I’ve sat with owners looking at a traffic report showing 400 visits in a month, and when I ask “what did you want those 400 people to do?” there’s a long pause. That pause is the real problem. The traffic was fine.

A shop with no till still gets foot traffic. It just never finds out who would have paid.

Different businesses, different conversions

Here’s where owners sometimes get tripped up: they look at what big brands or online stores measure and assume the same rules apply. They don’t. What counts as a conversion depends almost entirely on how your business actually gets paid.

Say you’re a plumber. The person who found you has water somewhere water shouldn’t be, and what they want is a human voice on the phone within the next two minutes. Your conversion is a phone call. Nothing else on your website comes close in value. An email enquiry from that same person is worth a fraction as much, because by the time you reply, someone else answered their call.

A therapist runs on a completely different clock. Nobody books therapy in a panic click at 11 p.m. People read the about page, close the tab, come back three days later, read it again, and then, maybe, request a consultation. That booked consult is the conversion, and the site’s job is to make the person feel safe enough to take that step when they’re finally ready.

A neighbourhood bakery might have the strangest conversion of all: the site never sees it. Someone checks whether you’re open Saturday morning, sees a photo of the sourdough, and shows up at 9 a.m. with cash. The website did its whole job and recorded nothing but a page view on the hours page. That’s still a conversion. It just happened at the counter.

And an online store is the textbook case. The sale happens right there in the checkout, on the site, with a number attached.

Even within one industry the answer shifts with the business model. A custom wedding cake studio converts through consultation requests, weeks ahead. The walk-in bakery two doors down converts through feet on the floor. Same flour, different job.

Primary and secondary conversions

Once you’ve named the main event, it helps to sort everything else your site could ask for into two buckets.

The primary conversion is the action closest to money. The call, the booking, the sale, the visit. One per page, and it gets the spotlight.

Secondary conversions are the smaller commitments for people who found you too early. They’re interested but they’re renovating in the spring, or they’re comparing three therapists, or they’re browsing your cakes at midnight for a wedding that’s fourteen months out. For those people, useful secondary asks include joining an email list, following your Instagram, downloading a pricing guide, or filling in a short “get a quote” form when the primary ask was a call. Each of those keeps a thread attached to someone who wasn’t ready today.

Two rules keep this from going sideways. First, the secondary ask exists to catch the not-yet crowd, so it should feel lighter than the primary one. Asking a first-time visitor to book a consultation and subscribe and follow you is how you get none of the three. Second, the secondary option never gets the loudest button. When I audit sites, I regularly find a giant colourful “Join our newsletter” banner sitting above a grey little “Contact us” link. That’s a page arguing with itself, and the newsletter is winning a fight it should never have been in.

Think of a restaurant host. The primary job is seating you. If the place is full, the fallback is taking your name for next time. The fallback matters, but no decent host leads with it.

“Get my name out there” is a wish, and a page can’t grant wishes

When I ask what a site should accomplish, the most common answer I hear from local business owners is some version of “I just want to get my name out there.” I understand the instinct, and as a business goal, awareness is real. As a job for a web page, it’s unusable.

A page can only cause behaviour. Clicks, taps, calls, scrolls, form fills. You can put a button on a page that says “Call now” and watch people press it. There’s no button for “become vaguely aware of us.” There’s also no way to measure it, and what can’t be measured can’t be improved. How would you ever know your name got sufficiently “out there”? What would that even look like on a Tuesday?

Here’s the move that unsticks this. Suppose the name-spreading worked perfectly. Someone in your neighbourhood now knows your business exists. Play the tape forward: what do they do with that knowledge? They search your name, land on your site, and then… whatever should happen next in that sentence, that’s your conversion. It was hiding inside the wish the whole time. Awareness was always a means to an action.

To be fair, simply having a decent website does carry weight on its own. The DreamHost 2026 Local Business Trust Index found that 69% of consumers consider a website essential for a local business to be credible, and 70% say they’re more likely to do business with a company that has one. So yes, existing online matters. But look at what that stat is really describing: credibility is the floor. It gets you considered. A conversion is what turns “considered” into “contacted,” and only one of those pays your rent.

How to pick the one action for each page

The whole site has a main job, and then each page has its own. Picking the per-page action comes down to two questions.

First: what’s the most valuable thing a visitor could do from this page? Value here means closeness to revenue. A phone call from a service page is usually worth more than a newsletter signup from the same page.

Second: what is this particular visitor actually ready for? This is the question people skip. Someone who landed on your pricing page is deep into a decision, so the big ask fits. Someone who arrived on a blog post about “how often should I service my furnace” is probably months from hiring anyone, and shoving a “Book now” ask in their face is like proposing on a first date. A smaller ask, like a maintenance checklist in exchange for an email, meets them where they are.

When the two questions disagree, my tiebreaker is this: pick the action closest to money that this visitor is plausibly ready to take. Homepage visitors searching your business by name? Ready. Give them the call. Blog readers researching a problem? Warmer to the small ask.

One caution before you go stripping pages down. “One job” doesn’t mean one link. Your navigation still exists, your footer still has the boring necessary stuff, and a visitor who wants your about page can still find it. One job means one ask gets the visual and verbal emphasis. Everything else stays available without competing.

And write your choices down. A job that lives only in your head has a way of drifting every time someone suggests adding “just one more thing” to the page.

Micro-conversions, and why the phone call runs local business

Between “visited” and “became a customer” there’s a trail of small footprints worth knowing about. Marketers call them micro-conversions: tapping your phone number, looking up directions, checking your hours, submitting a short form, clicking through from your Google Business Profile. Individually tiny, collectively they’re the clearest evidence of intent your site produces.

They matter most when your final sale happens away from the website. The bakery from earlier can’t track cash at the counter through analytics, but it can count direction lookups and hours checks, and those numbers rise and fall with the till. If direction lookups doubled after a site refresh, the refresh worked, even though no software recorded a single “sale.”

Among all of these, for local service businesses, the phone call towers over the rest. A few reasons stack up:

• Urgency. The person with a burst pipe, a dead furnace in January, or a car that won’t start on the Gardiner isn’t filling in a form and waiting to hear back. They’re calling whoever answers first.
• Specificity. Real jobs resist form fields. Try describing the exact grinding noise your dryer makes in a 200-character message box. On the phone it takes eight seconds and an impression.
• Trust. Before letting a stranger into their house, people want to hear a voice. Thirty seconds of a calm, competent human on the line does more than any page of testimonials.
• Speed to yes. A call can end with a booked appointment. A form submission ends with waiting.

If you run a trade or any urgent service, this settles the priority order. Design for the call first. Let the form be the backup for the 2 a.m. crowd and the phone-shy, and make sure it promises a response time so they know what waiting means.

The action you pick changes the whole design

This is where the decision stops being philosophy and starts moving pixels. Watch what happens to the same five-page site under different jobs.

If the job is a phone call, the number lives in the header of every page, large and tappable. Your hours sit right beside it, because a number with no hours is an invitation to play voicemail roulette. Buttons say “Call 416-555-0142” rather than “Contact us,” because naming the action removes a decision. The contact form shrinks to a fallback with a promised reply time.

If the job is a booked consultation, the booking link or embedded calendar becomes the star, and everything around it works to lower the nerves. A therapist’s site should answer, before the ask, what a first session involves, how long it runs, and what it costs, because uncertainty is exactly what stops that particular visitor. Every screen of reassurance earns the button that follows it.

If the job is getting people through your door, the hours go above the fold, the map gets embedded rather than linked, there’s a line about parking or the nearest streetcar stop, and the photos show your actual counter and your actual croissants. Holiday hours get updated the week before the holiday, because “shows open, was closed” burns trust you won’t get back.

If the job is an online sale, the weight shifts to product photography, shipping costs surfaced before checkout instead of springing out at the end, a returns policy in plain sight, and the fewest possible steps between cart and paid.

Same business category, four different websites, and every difference traces back to the chosen action. This is also why “can you just make it look more modern” is an unanswerable brief. Between two designs, prettier is a matter of taste. Better at causing the chosen action is a matter of fact, and you can only get to facts after you’ve chosen.

A ten-minute exercise: give every page its one job

Here’s something you can do tonight with a pen and the back of an envelope.

List every page on your site. For most local businesses that’s five or six lines: home, services, about, contact, maybe a gallery or a blog. Next to each one, finish this sentence: “When someone lands here, I want them to ___.”

A few rules to keep it useful:

1. One verb per page. If you wrote “and,” start over on that line.
2. It has to be something a person can physically do during the visit. “Call us” counts. “See that we’re professional” doesn’t, because seeing leads nowhere on its own. If professionalism is the point, the line becomes “trust us enough to call,” and the verb is still call.
3. If you truly can’t finish the sentence for a page, circle it. A page with no completable sentence is either dead weight or a job you haven’t discovered yet.

Then open each page next to your list and check one thing: does the most prominent element on that page point at the verb you wrote? Prominent means what your eye hits first, the biggest button, the thing above the fold on a phone.

Typical findings from doing this with real sites: two pages sharing the same job (candidates for merging), a services page whose sentence ends abruptly because there’s no ask on it anywhere, and a homepage trying to perform all five jobs at once, which reads like five people talking over each other.

Keep the envelope. It’s now the shortest useful design brief in existence, and it’ll outperform most twenty-page ones. Hand it to whoever touches your website next and they’ll know exactly what each page is being graded on.

Want a site where every page already knows its job?

This exercise is the actual starting point of my build process. Before I write a headline or pick a font for a client, we settle what each of the five pages is supposed to cause, and every decision afterward answers to that list.

If you’d rather start there with me than with the envelope, the offer is simple. My standard build is $1,000, and new clients currently pay $500. That covers the full five-page site with the first year of hosting, updates, and support included, then $250 a year afterward if you’d like me to keep looking after it. Everything is laid out at hughmcnamara.com/new-clients.

Tell me what your business does and how customers usually reach you, and before we talk money I’ll tell you what I think your website’s one job should be. Worst case, you walk away with a clearer envelope.