LOCAL SEO CASE STUDY, WITH DATES
A local SEO case study covering a three month, part time engagement for a Bay Area luxury transport company, then a deliberate pause. The interesting part is not the growth during the sprint. It is what the rankings did for the twelvt months after I stopped touching them.
One client. One industry. Read the limitations section before you draw conclusions.
THE ENGAGEMENT
Every figure on this page carries the account it came from and the date range it covers. The screenshots are below, so you can check each one rather than take my word for it.
The limo marketing playbookFourteen months, and I worked on it part time for three of them. Impressions went from nothing to 42,800, which means Google now puts this site in front of people. Average position improved from 48.8 to 32.2, sixteen places, and the click through rate went from 0.5 to 0.8 percent. Both of those moved in the right direction during months when nobody was touching the site.
Here is the query report, which is the part almost nobody publishes. Google now shows this site for the searches that actually buy: 174 times for “black car service san francisco”, 158 for “luxury black sedan limo san francisco”, 125 for “black car service san jose”. Two years ago it appeared for none of them. That is the foundation, and it is real.
And every one of those three earned zero clicks. Nine commercial searches carried 1,016 impressions between them and produced nothing. Five of the eight clicks in the report came from people typing the company name, who were going to find it anyway. So the honest reading is that Google now understands what this business is and still ranks it too low for anyone to click, which is a different problem from being invisible and a much better one to have.
Now the rest of what most agencies would leave out. The site-wide average position is still 32, page three or four across everything it appears for, and 195 clicks across fourteen months is a small number. Ranking is not the same as being found, and being found is not the same as being booked. What this proves is that the foundation holds and keeps compounding on its own. It has not yet made this business the first result for a competitive term, and five months of work was never going to.
That is the honest state of this engagement, and the timescale is the part worth holding on to. Three months of part time work, then a year of nothing, and it was still climbing at the end of that year. It has not made this business the top result for a competitive term, and three months part time was never going to. What it did was build something that kept working after the invoicing stopped.
THE ACTUAL SCREENS
Two screens, straight from the client's Google Search Console. The first covers all fourteen months. The second is where it stood eight months after I stopped. The Semrush keyword report, the indexing report and the intent map come from the same account and I will show you those live rather than stack seven screenshots on a page.
Note for the site owner: the two videos on the current live case study page, client-previous-website-.mp4 and Client-website-transition.mp4, both return 404. The before and after of the old website is missing from the site. If you still have those files, re-upload them and I will add a proper before and after block here.
THE OTHER THING ON THIS PAGE
Everything above was done by hand: a technical rebuild, keyword mapping, titles rewritten one at a time. That is the SEO service. It is not the automation. The automation is a different product, and this is the only account I can show you running it, because the account is mine.
Yes, I am publishing my own bad number on purpose. Two clicks from twenty thousand appearances is not a brag, it is a diagnosis, and I can name it exactly: position 14. Page two. The same diagnosis I gave the limo company, visible to Google and still too low to be reached, except mine has a number on it and four or five places to go.
Consider who is above me. Page one for that term is agencies with twenty year old domains who sell this for a living. Getting to fourteen from nothing, on a two year old site, against that, is the part I would ask you to judge. Not the click count.
Your market is almost certainly easier than mine. A day care, a limo firm or a dental practice is not competing against an industry of specialists for its own keywords. That is the honest reason this work tends to move faster for you than it has for me, and it is a better argument than a screenshot of somebody else’s good quarter. I am the client in this section, so weigh it less than the account above, which belongs to someone else and which I will open live on a call. Read something the software wrote ↗
THE OTHER HALF OF THE WORK
Not every business is losing to search. Some are losing because nobody has seen what it is like to walk in, or because the only picture of the owner is a cropped phone photo. These are samples of the work, not case studies. There is no ranking chart under either one, and for this kind of work there should not be: the film and the photograph are the evidence. You can judge them in about four seconds, which is the same four seconds your customer gets.
Video. Shot and cut by me for a day care in Walnut Creek. A production sample: I have not run a full engagement in that niche yet, so there is no chart to put beside it and I am not going to invent one.
Straight about what these prove. They prove the standard of the work, which for a film and a portrait is the whole question. They do not prove a ranking, a booking or a revenue figure, and I have not dressed them up as though they do. The one engagement on this page with numbers attached is the limo company above, and its numbers are shown whole, including the unflattering half.
WHAT ACTUALLY HAPPENED
The pause was planned. A foundation you cannot stop paying for is not a foundation, it is a subscription.
The site existed and Google largely was not storing it. There was no Search Console, no Analytics, and the page structure gave a crawler nothing to work with. Semrush showed the organic keyword count as a flat line near zero. For a business where customers search a service plus an airport or a city, that is close to being invisible.
Fixed the crawl errors, rebuilt the page structure and hierarchy, connected Search Console and Analytics so there was finally a measurement baseline, and brought the mobile experience to parity with desktop.
Mapped the high intent luxury transport searches to specific pages, one intent per page rather than one page trying to rank for everything. Rewrote titles and descriptions per page and built the internal links so a new page inherited authority rather than sitting alone. This is the period where the impression curve starts climbing.
No retainer, no maintenance, no ad spend, no new content. The question we both wanted answered was whether the structure would hold on its own or whether the results were rented.
The highest click days sit just after the last change shipped. That lag is normal and worth understanding: Google reprocesses a rebuilt site over weeks, so the month you stop paying is rarely the month results stop arriving.
Impressions softened through September and October. This is what happens to any site with no fresh content and no new links, and I am including it because a case study that only shows the line going up is an advertisement, not evidence.
Over those three months Search Console shows average position 32.2 against an all time average of 48.8, and click through 0.8 percent against 0.5 percent. Better placement, better click rate, zero spend. That is the actual finding, and it is narrower than it sounds: the structural work is an asset the business owns, but the site is still ranking on page three for its competitive terms, not page one.
42,800 impressions and 195 clicks are the same story told twice.
Rex hosts the measurement, because a number without its denominator is a decoration. Impressions say Google understands the site. Clicks say it is still ranking too low for anyone to press.THE STATE OF THE SITE
Not a design opinion. These are the specific conditions that decide whether a page can rank at all.
READ THIS PART
I would rather you trust the small claim than doubt a big one. Here is the honest scope of what a single engagement can tell you.
How I handle claims and data ↗One business, one industry, one market. It is not a sample. It is evidence that this approach worked once, with dates you can hold me to.
Competition, search volume and how much a customer is worth all differ enormously by category. A dentist in Oakland and a limo service face different problems.
Search Console shows clicks and positions. It does not show bookings. I am reporting what I can actually measure rather than a revenue figure I would be guessing at.
The site held because the structural work was sound, and it still declined before recovering. A site with ongoing content would not have had the dip at all.
THE SAME WORK, FOR YOU
Three things in a fixed order. It is the same order I would run for your business, and the audit decides how much of each you actually need.
Indexing errors, page structure, titles and descriptions, mobile performance. Until this is done nothing else you pay for can work, which is why it comes first every time.
Technical SEO ↗Find the searches with real local intent, then give each one a page built to answer it. Most sites fail here by asking one page to rank for eight different things.
Local SEO service ↗Fresh content, profile activity and new links. The dip in autumn 2025 is exactly what skipping this looks like. Automating it is what the engine exists for.
The automation engine ↗FAIR QUESTIONS
Because I only have one engagement with a full dated record and Search Console access I can still pull from. I could pad this page with vague client logos and unattributed quotes. Every agency does. You would learn nothing from it and I would have taught you that my numbers are decorative.
Almost certainly not, and anyone telling you otherwise is guessing. Your competition, your category, your domain age and your budget are all different. What transfers is the method and the order of operations, not the outcome.
They are exports from the client's own Google Search Console and Semrush accounts, and the date range is visible in each one. Ask at the consultation and I will open the live account on the call so you can watch the numbers load rather than trust an image.
Correct, and I would rather say so than hide it behind a percentage. Two things explain it. The site started from almost nothing, so the growth is real but the base was tiny. And average position is 32, which is page three, where almost nobody clicks. The impressions prove Google now understands what the site is for. Turning impressions into clicks is the next job and it needs content and links, which is exactly the work that stopped in July 2025.
Sometimes, and that is a real answer rather than a sales one. Technical and structural work is genuinely permanent. Content and links are not, and the autumn dip on this timeline is what stopping looks like. If your market is competitive, someone will out publish you eventually.
Run the free checker for the technical picture, which costs nothing and takes a minute. For the full version, the consultation produces the same kind of written analysis and prioritized plan that started this engagement.
RELATED PAGES
KEEP EXPLORING
The same principles, written for a business doing it without help.
Read the article ↗
AUTOMATED MARKETING
SEASONAL MARKETINGYOUR SITE HAS ITS OWN VERSION OF THIS TIMELINE.
Morning and evening editions
A free, short email for Bay Area small business owners: the business news that matters, and what to do about it.
Check your inbox for the welcome email.