Your homepage is not a landing page

Almost every small business runs ads to the homepage. It is the URL everyone knows, it is the page that got the most attention during the build, and pointing there feels safe.

It is the most expensive default in local advertising. You paid for a click from someone searching for one specific service, and you delivered them to a page whose main job is explaining that your company exists.

What is the difference between a homepage and a landing page?

The number of jobs.

Your homepage has to work for a customer looking up your hours, a stranger deciding whether you are real, someone hunting for the careers page, a supplier looking for accounts, and a search engine trying to understand your business. Serving all of them means compromising for each of them.

A landing page has exactly one job. Someone searched for emergency drain clearing in your city, clicked your ad, and needs to know within seconds that you do emergency drain clearing in that city and how fast you can get there. Everything else on the page is a distraction competing with the phone number.

That single-purpose focus is why landing pages convert better. Not design, not persuasion tricks. Fewer exits and one clear next step.

A homepage answers "who are you." A landing page answers "can you do this specific thing for me right now."

Why does sending ads to the homepage cost money?

Because the visitor arrives with a question and has to go find the answer themselves.

They searched for one service. Your homepage lists eleven. Now they are scanning a services menu, deciding which link matches what they typed, and every extra step is a chance to abandon. You already paid for that click either way.

Put a price on the abandonment. LocaliQ and WordStream's 2026 search advertising benchmarks put the median cost per click across all industries at $5.42, rising to $8.33 for Home and Home Improvement and $8.00 for dentists. Those are medians, from a vendor that sells advertising services, and the 2026 sample size is not published. At $8.33, every visitor who lands on your homepage and cannot find their service in ten seconds cost you the price of a coffee, and you will never see them again.

There is a second cost that owners underestimate. Homepages are full of navigation, and navigation is a set of exits. Every menu item, every footer link, every "learn more about our story" is a route away from the action you paid for.

And the message does not match. Your ad promised same-day service, the homepage headline talks about family values and craftsmanship. That mismatch reads as a bait and switch even when the business is completely honest, and the visitor leaves without being able to say why.

The usual response to poor ad results is to cut spend, but a mismatched destination is a page problem, not a bidding problem. It is worth reading the case for pausing a campaign rather than optimizing it before you spend another month tuning bids on traffic that lands somewhere wrong.

How urgent is a local searcher, really?

Urgent, but not as urgent as the pitch says, and the difference matters because the whole case for a stripped-down landing page rests on this claim.

You will be told that 88% of local searches result in a call or a visit, or that 76% of local searchers visit a business within a day. Neither figure appears in any primary document. Both are corruptions of one piece of research, and it is worth knowing what that research actually said.

The original is a May 2014 study Google ran with Ipsos MediaCT and Purchased. It is two studies stitched together: an Ipsos online survey of 4,500 respondents across nine verticals, fielded 10 to 22 January 2014, and a Purchased mobile diary in which 3,431 smartphone searches and 2,262 store visits were logged over seven days between December 2013 and January 2014. The deck's actual sentence is that 50% of consumers who conducted a local search on their smartphone visited a store within a day, and 34% who searched on a computer or tablet did the same. It also reports that 18% of local searches on a smartphone led to a purchase within a day, against 7% for non-local searches.

Two things follow. First, half is a genuinely strong number and it does support building a page that lets someone act immediately. Second, it is 2013 and 2014 fieldwork, older than the current local pack, older than local pack ads, older than AI results. Anyone quoting 88% is quoting a number that has been round-tripped through enough marketing blogs to lose its source entirely, and anyone quoting the real 50% without the date is only half honest.

Build the landing page. Just build it because a single-purpose page has fewer ways to fail, not because a twelve year old statistic promises a same-day visit.

What belongs on a landing page?

Less than you think, arranged in a specific order.

A headline that repeats the words in the ad and names the city. If the ad said emergency water heater repair in Oakland, the headline says that. Matching language is the whole trick.

The action, immediately. A tappable phone number or a short form, visible without scrolling. Then repeated further down for the people who read everything.

Proof that survives scrutiny. Review count, license number, how long you have been doing this, real photos of real work. Trust elements that change behavior are specific and checkable, and a landing page is where the generic ones do the least good.

What happens next, in plain terms. You call, we answer, a technician arrives in a window we give you. Reducing uncertainty converts better than adjectives.

And almost no navigation. A logo, the phone number, and the content. Nothing else needs to be clickable.

Two technical points matter more here than elsewhere. The page needs to be quick, because paid visitors have no loyalty and speed matters exactly where people are still waiting. Google's current Core Web Vitals thresholds put Largest Contentful Paint at 2.5 seconds or less for a good rating and worse than 4 seconds for poor, assessed at the 75th percentile of visits, and on a landing page the element deciding that is almost always the hero image. And it needs to be usable by everyone, since accessibility basics improve conversion for the same reasons they improve access: larger tap targets, real labels, readable contrast.

Do I need one page per campaign?

One page per distinct promise, not one per campaign.

If you run six ad groups that all say roof repair with slightly different keywords, one strong roof repair page serves all of them. If one campaign sells roof repair and another sells full replacement, those are two different buyers with different urgency and different price expectations, and they deserve two pages.

Start with your highest-spend service. Build one page, run it against the homepage for a quarter, and compare the cost per real lead, not per click.

The part that goes wrong most often is measurement. New pages mean new forms and new events, and tracking silently breaks. Build it so that your conversion tracking survives a website change, or you will end up comparing a page that counts leads against one that does not.

How long before I can tell which page won?

Longer than the platform will let you believe, and the honest answer changes what you should measure.

Run the arithmetic on a realistic budget. Say $1,500 a month in a home services category, at the $8.33 median cost per click. That is about 180 clicks a month, 540 over a quarter, split evenly between the homepage and the new landing page. Each arm gets 270 clicks.

At the 8.05% median conversion rate for that category, 270 clicks produces about 22 leads per arm across the whole quarter.

Now compare that to what a test would actually require, which is set by the effect you want to catch and not by a visitor count. Conversions needed per variant is roughly 16 ÷ r², where r is the relative lift and the 16 is 2(1.96 + 0.84)² rounded, 1.96 for 95% confidence and 0.84 for 80% power. A hundred conversions per arm is exactly what detects a 40% lift (derived), and no page swap moves booked work by two fifths. At 22 leads per arm the smallest difference you could resolve is about 85%derived. You are not running an experiment. You are running a comparison, and the honest report says so.

So measure the thing that survives small numbers. Take cost per booked job over the full quarter, not cost per click and not cost per lead. At 22 leads per arm and a one in three close rate, each arm books about seven jobs on $2,250 of spend, which is $321 a booked job. If the landing page arm books ten instead of seven, that is $225 a booked job, and a difference that size is worth acting on even though nobody can call it significant. A difference of one job is not.

Two rules keep this defensible. Run both arms for the same full weeks, because splitting a quarter unevenly imports seasonality straight into the result. And do not change either page mid-flight, no matter how obvious the improvement looks in week two.

What to do this week

Open the ad you spend the most on. Click it as a customer would, on a phone. Read the first screen you land on and ask whether it repeats the promise in the ad.

If it does not, write the headline for a dedicated page this week. One service, one city, one action. You can build the rest of the page later, but the headline is the decision.

Then count the exit links on your current destination page. Menu items, footer links, social icons. If a paid visitor has more than a handful of ways to leave before acting, that number is your first edit.

Work out your own leads per arm per quarter using the arithmetic above, before you promise anyone a test. Knowing you have 22 per arm rather than the hundred derived above changes how you write the report and stops you claiming a winner you cannot support.

Last, confirm the form on that page fires a conversion event. Submit a test and check it lands in analytics, not just in your inbox.

Questions about where your ads are landing?

Email me at eric@seod.com with the text of your best-performing ad and the URL it points to. I will tell you whether the promise and the destination match, and if they do not, I will write the headline that would fix it and send it back to you.

It takes me a few minutes and there is nothing attached to it. If the answer is that your current page is fine and your problem is targeting, I will say that instead.

There is also more on conversion in the library.