The Click Is Disappearing. The Decision Is Not.

For years, online marketing had a simple shape. Someone searched, found your business, clicked your website, and called. Analytics recorded the path. Marketing got the credit.

That path still exists. It just explains less than it used to.

In U.S. clickstream data, 58.5% of Google searches in 2024 ended without a click. For every 1,000 Google searches, only 360 clicks reached a website outside Google’s own properties and ads. ¹

Search is not disappearing. People are searching constantly. The change is what happens between the question and the website.

For contractors, that changes what it means to be found.

Search is no longer one place

Google still dominates search. In a study of search activity across 41 major websites in late 2025, Google alone accounted for 73.7% of U.S. desktop searches. Traditional search engines accounted for about 80% overall. The rest happened across commerce sites, social networks, AI tools, and other platforms. ²

A homeowner with a failing air conditioner might search Google, open Maps, read reviews, ask a neighborhood Facebook group, watch a YouTube video, or ask an AI assistant what to look for in an HVAC company.

They may encounter the same contractor several times.

Then they search the company name, visit its website, and call.

Analytics sees Google.

The customer journey was larger than Google.

This distinction matters because businesses have spent years treating the measurable visit as the beginning of the decision. Increasingly, it may be near the end.

Your website may not make the first impression

A search result can introduce the company. So can a Google Business Profile, a review, a project photo, a recommendation, a social post, or an AI answer.

Your website has not become less important. It has become less likely to be the first thing someone sees.

That gives it a different job.

When a prospective customer reaches the site, they may already know the company name. They may already know the review score. They may already have heard that you install mini splits, work in their town, offer financing, or answer emergency calls.

Now they are checking.

Is this the right company? Do they actually do this work? Do they serve my area? Do they look credible? Can I trust them with a $12,000 replacement?

The website has to answer those questions, but it has another audience now too. Search engines and AI systems use public information to understand what a business is, what it does, where it works, and whether it is relevant to a question.

That makes the public representation of the business more important, not less.

Attribution tells you what received credit

There is another problem with the old model. We can see less of the customer journey than our dashboards imply.

In one controlled experiment, researchers sent 1,113 visits across 11 social networks and messaging platforms, then compared the known source with what Google Analytics reported. Every visit from TikTok, Slack, Discord, WhatsApp, and Mastodon was classified as direct traffic. Facebook Messenger hid the source on 75% of visits. Instagram messages hid it 30% of the time. ³

The traffic was real. The source disappeared.

That is the easier attribution problem.

The harder one is deciding whether the channel that received credit actually caused the sale.

Dropbox tested this at a much larger scale. Its data science team temporarily shut off entire advertising channels and compared the resulting business performance with what normal attribution had claimed. Mobile advertising showed an attributed return on ad spend of 1.53. Its experimentally measured return was 0.70. Search advertising appeared to return roughly 2.0 through attribution. The causal estimate was 0.92. ⁴

Dropbox eventually moved about $25 million away from lower-incrementality advertising, while its incrementality-adjusted lifetime-value-to-acquisition-cost ratio improved 53%. ⁴

A contractor does not need a Dropbox-sized marketing department to understand the lesson.

The channel that recorded the customer is not necessarily the reason the customer chose you.

A neighbor may have recommended you. Your truck may have been parked three houses down. They may have read six reviews. They may have seen your name in Maps twice. Then they Googled the company and called.

Organic search gets the conversion.

That does not mean organic search created the demand.

Analytics is still useful. It just describes the evidence it can see.

Being found now means being represented accurately

This is where the change becomes practical.

Imagine an HVAC company that installs heat pumps every week, serves six surrounding communities, handles commercial rooftop units, offers emergency service, and has hundreds of completed projects.

Now imagine its website has five pages.

Home. About. Services. Contact. Financing.

The company and its public representation are two different things.

A customer looking specifically for heat-pump replacement may not find much evidence. Google has less information to connect the company with that work. An AI system has less useful material to cite or summarize. Someone trying to determine whether the company serves a nearby town may have to guess.

That is a coverage problem.

The answer is not to publish hundreds of pages.

A keyword is not a reason to create a page. A city name is not a reason to create a page. Search volume alone is not a reason to create a page.

Something real has to exist behind it.

A real service. A real market. A real customer need. A real offer. Useful proof. Information that deserves a permanent place in the public record.

That is how we think about Search Coverage at Rule & Form.

The question is not simply whether another page could rank.

The question is whether the business is being represented where it legitimately should be.

More traffic is not the final score

Rankings matter. Traffic matters. Impressions matter.

They are diagnostic signals.

A contractor ultimately needs qualified inquiries, booked estimates, jobs, repeat customers, and profitable work. Marketing should be judged increasingly close to those outcomes.

That does not mean abandoning search metrics. It means putting them in their proper place.

If impressions rise while clicks fall, something may have changed in the search results. If branded searches rise, awareness may be increasing. If a new service page begins appearing for valuable searches, coverage may be improving. If leads increase but booked work does not, the problem may be somewhere else entirely.

No single number tells the whole story.

That is uncomfortable after twenty years of digital marketing promising that everything could be attributed cleanly. It is also closer to how customers have always made decisions.

They hear about companies. They notice them. They compare them. They ask people they trust. They check reviews. They search. They come back later. Eventually they choose.

The internet is getting better at hiding the middle of that process.

The businesses that adapt will make sure the important facts about them are easy to find anyway. They will build a stronger public record, cover the work and markets that genuinely belong to them, and measure success as close to actual business outcomes as they can.

The click may be disappearing.

The decision is not.

Sources

1. SparkToro and Datos, 2024 Zero-Click Search Study. The study reports that 58.5% of U.S. Google searches ended without a click and that 360 of every 1,000 searches produced a click to the open web.
2. SparkToro and Datos, Search Happens Everywhere, based on U.S. desktop clickstream activity across 41 websites in Q4 2025.
3. SparkToro and Really Good Data, Dark Social Falsely Attributes Significant Percentages of Web Traffic as Direct. The experiment recorded 1,113 visits over ten days across 11 networks.
4. Varun V. Chivukula, Ling Jin, and Julia Zhan, From Attribution to Causality in Digital Advertising, IEEE Access, 2026.