Pick up your phone and type your own town plus flooring into Google. Look at what comes back before you scroll.
An ad or two. A small map with three companies and their star ratings. Then the blue links. And near the top, a paragraph of AI written text answering the question so nobody has to click at all.
That screen is the whole job. Everything sold to you under the word SEO happens on it, and the three pieces do not send you the same kind of person.
The three surfaces, and what each one really sends you
| Surface | What it actually is | What it sends a flooring store |
|---|---|---|
| The map pack | Three Google Business Profiles, ranked mostly on how close you are and how complete the profile is | Calls and direction taps from people shopping for somebody to come out. The closest thing on the page to a booked estimate. |
| Organic, the blue links | Pages on your own website | Research traffic. Cost questions, product comparisons, people three weeks out. Real, and slower. |
| The AI answer | A model reading pages, profiles and reviews, then writing a paragraph that sometimes names companies | A few people who arrive knowing your name. Nobody can honestly size it yet. |
If you sell and install floors in a defined area, the map pack is where the estimates are. Somebody typing "flooring installers near me" into a phone at nine in the morning is not researching. They want a person to come measure. Different human being from the one reading about engineered versus solid.
The other two are not optional. The map points at your website, and the AI answer reads your pages and reviews to decide what to say about you. It is one job. Start with the profile, because it pays back first.
The profile, and the decision people keep flip flopping on
You pick one primary category and a stack of secondary ones. The primary carries the most weight, and it is the field contractors get told two different things about by two different advisors. Google's list has both a flooring store entry and a flooring contractor entry, and there is no universal answer, because the field describes your business rather than tricking anything. Pick the one a customer would use for you.
- You have a showroom. People walk in, handle samples and buy material. Your primary is the store one, because that is what you are.
- You have no showroom. You measure, you install, you drive to them. Your primary is the contractor one. Filing as a store you do not have puts you beside stores you cannot beat on the one thing store shoppers came for.
Then add everything else as a secondary. Hardwood floor refinisher, carpet installer, tile contractor, whatever you genuinely do. They cost nothing and they are how you turn up for the specific job.
What costs money is swapping the primary back and forth. Every swap throws away whatever the profile had earned in the categories it was already showing up for. Pick the true one and leave it.
Two other fields. The business name has to be your real name, not your name with keywords stapled on, because competitors report those. And the address is the one you cannot argue with. A profile ranks strongest near where it sits, so if the towns you want are forty minutes out, the map will not reach them. That is what the website section below is for.
What actually moves a profile, in rough order
- Where you are. Not adjustable, and the biggest single input in the map. A store on the far edge of a metro cannot rank in the downtown pack no matter who it hires.
- Categories. Primary correct, secondaries complete. Twenty minutes, and most profiles carry three where they could carry nine.
- Reviews, and the pattern more than the pile. A range with reasoning: two to four a month, every month, beats thirty in one week and nothing for a year. A steady trickle reads as a company working right now. A pile with a gap behind it reads as one review push. And last week's review is the first one a homeowner reads.
- Services, written out. Every service as its own entry, in the words a customer says rather than the words a supplier says. Free, quick, and skipped on most profiles here.
- Photos. A range with reasoning: ten to twenty real job photos in the first month, then a handful monthly. The reasoning is that a homeowner reads these to decide whether you are a real company, and four stock images of a laminate plank do not clear that bar. The subfloor. The truck. The finished room from the doorway.
- Posts and the questions section. They do not move rank much. They keep the profile current and let you answer the price question in public. Check the message setting while you are in there: if the profile says you reply and nobody does, that is an inquiry dying where you never look.
All facts about your own business, and none of it needs a vendor. The deeper walkthrough is local SEO for a flooring company.
An illustration: what one position in the map is worth
Round invented numbers. Not our results, not a projection for you. Arithmetic, so you can drop your own figures in the same slots. Step one is a number you already have, because the profile reports its own calls.
- At the position you sit today, the profile produces 20 calls a month.
- The work above moves you up the pack and it produces 35. Fifteen more calls.
- Two thirds are real inquiries rather than wrong numbers. Call it 10.
- Half of those book an estimate. Five more estimates a month.
- You close one in three. One or two more jobs a month.
- At a $6,000 average job, near $9,000 a month in sold work, off an asset you own.
It runs downward too. If nobody answers the phone inside a few minutes, step three is where the chain dies, and you paid for a better position so you could send more people to voicemail. The number to judge this by was never a ranking. It is what a booked estimate costs you.
City pages, done honestly and done stupidly
The map is bound by where you sit. Your website is not, and that is the reason city pages exist. It is how a store in one town turns up in the four towns it drives to.
It is also the easiest thing here to do badly, and the bad version has a name. A doorway page is one template repeated for forty towns with the name swapped in. Google has been explicit about that pattern for years, which makes it one of the few things here that is not opinion.
| A real city page has | A doorway page has |
|---|---|
| Jobs you really did there, with your own photos | Stock photos, or the same photos on all forty pages |
| Neighborhoods by name, and what the housing stock means for the work. Seventies builds on that side of town, so plywood subfloor and a lot of tile removal | The town name pasted into a sentence that would be true anywhere in the country |
| Your real drive time and your real radius | Forty towns, several you have never worked in |
| One page per town, added as you serve it | Forty pages published in one week |
So write one. Good enough that somebody who lives there reads it and thinks, they know this place. Write the second a month later. If you cannot write a real page about a town, you do not serve it yet. And for most stores the pages that pay off faster are not city pages at all. They are the job plus the place, and the question pages. Build that list before you write anything.
What the AI answers change, and what they do not
Plain version. A model reads pages, profiles, reviews and directory entries, then writes a paragraph answering the question. Sometimes it names three companies. It is a fast version of what a homeowner used to do with six tabs open. Three things follow.
- It reads text, not pictures of text. And not text that only appears after a browser runs a script. Open your own site with script turned off. If the words are gone, some of what reads the web sees the same empty page, and that is the job before every other one here.
- It reads your profile and your reviews too. The profile work above is not a separate project from this one.
- It rewards a page that answers one question completely. A page saying what a floor costs in your area, as a range, with the reasoning beside it, is the kind that gets quoted. A page of adjectives is not.
Now the part nobody selling this will say. Nobody, us included, can tell you what share of your calls came through an AI answer, because the reporting is not there yet. Anybody handing you an AI search package with a number attached invented it. Do the three things, watch your phone, stop there.
What to do yourself, and what to pay for
| The job | Who does it, and why |
|---|---|
| Profile category, services, hours | You. Facts about your own business. Two hours, once. |
| Asking for reviews, and replying | You, on the job. The ask that works comes from the person who did the work, in the finished room, on the day. |
| Photos | Your crew. Five minutes at the end of a job. |
| Site structure, speed, readable text | Pay. A build, not a habit, and expensive to undo wrong. |
| City and product pages | Pay for the writing, supply the truth. The jobs, photos and neighborhoods can only come from you. |
| Directory listings | Pay once, not monthly. One cleanup so your name, address and phone match everywhere. |
| Links from other sites | Careful. The safe version is local and boring: suppliers, associations, the team you sponsor. Bought packages of links can hurt you. |
How long it honestly takes
Ranges, with the reasoning that produces each one. Not a guarantee. Anybody giving you a date is guessing at a system they do not control.
| The work | First real movement | Why it lands there |
|---|---|---|
| Profile fixes: category, services, photos, hours | 2 to 8 weeks | Nothing has to be earned. You are supplying facts that were missing and the map re-reads them fast. |
| Reviews moving where you sit in the pack | 3 to 6 months | You have to do the jobs first. At two to four a month, a pattern needs a couple of quarters to look like one. |
| One city page ranking in its own town | 3 to 9 months | It has to be found, then outlast whoever sits there now. Small towns land short, because there is less to beat. Metro suburbs land long. |
| Competitive product terms across a large metro | 9 to 18 months, sometimes never | You are up against national retailers and manufacturer sites that have held those pages a decade. Often the money belongs elsewhere. |
A promise of movement in thirty days is either profile work sold under the word SEO, which is fine as long as it is priced that way, or it is nothing.
The mistake that wastes the most money
It is not technical. It is paying every month for movement on words nobody hires from.
Twelve months in, the report is green. Forty keywords up. A chart with an arrow. And the phone rings as often as it did last year, because the words that moved were "hardwood flooring types" and "best underlayment" and "how to fix a squeaky floor board". Not one of those people is hiring anybody.
The fix costs nothing and it happens before you sign. Write down 10 to 15 searches. Real ones, the ones you would type if your own floor needed doing, with your real town names inside them. Hand the list over and say: report on these. A year later, a report showing where you sit on each one tells you whether you bought anything. A list that gets refused, or quietly swapped for a longer one, told you something for free. Put it next to the questions worth asking any agency.
The second most expensive habit is starting over. New site, new agency, new primary category, every nine months. Every restart throws away the one thing here that compounds.
The shortest version
The map pack is where the estimates are, and the profile that wins it is mostly facts you can enter yourself this week. Your website reaches the towns you drive to, one honest page at a time. The AI answers reward those same plain readable pages, and nobody can size them yet. Judge all of it on what a booked estimate costs you, never on a chart of words.
