How we saved $30K in 16 months and got 76.6K visitors for Cevik Motors

Cevik Motors (Çevik Motorlu Araçlar) sells salvage and second-hand vehicle parts out of Türkiye. When they came to us, every single customer was bought. Google Ads at roughly $0.50 a click was the entire acquisition channel, and it behaved exactly the way rented traffic behaves: the phone rang while the budget ran, and went quiet the day it stopped.

We proposed the opposite trade. Turn the ads off completely, put the whole budget behind search, and build a catalogue Google can rank on its own. That is a genuinely uncomfortable ask for a business whose leads all come from one channel, and they took it. This is what the sixteen months after that decision look like.

The problem with buying every click

Paid search has no residual value. A click costs the same on day one and on day four hundred, and nothing you buy compounds. For a parts business that is particularly punishing, because the catalogue is enormous and the searches are long-tail: somebody types the exact part, for the exact make, for the exact model year. Bidding on that many keyword combinations is expensive and most of them are too low-volume to be worth a campaign at all.

But long-tail is exactly what organic search is good at. Those searches are cheap to rank for, they carry buying intent that is almost embarrassing in how direct it is, and once a page ranks it keeps working. The catalogue that made paid search unaffordable is the same catalogue that makes SEO work.

What we actually built

Nobody writes fifty thousand pages by hand. The work was programmatic: take the structured data the business already has, which is stock, and turn it into real, crawlable URLs, each one aimed at a search someone is genuinely typing.

• A page per part category × make × model, generated from their own inventory rather than authored one at a time
• Category pages that match how a Turkish buyer actually phrases it, e.g. "çıkma parça Volvo", not how a parts database names it
• Technical foundations before content: indexable server-rendered HTML, clean URL structure, internal links that actually connect the tree, and pages fast enough to survive on mobile data
• Index management: getting the pages that matter crawled, and keeping the thin ones from eating the crawl budget, which is where most programmatic projects quietly fail
• A phone number and a WhatsApp link on every one of those pages, because in this market the conversion is a call, not a cart

The walkthrough

I recorded the method rather than just describing it: how the pages get generated, how they get indexed at volume, and what the indexing curve looks like when it works.

The 8-month checkpoint

We published the numbers publicly at the halfway point, so this is not a figure reconstructed after the fact. In October 2025, eight months in, the same property had done 31.8K clicks and 1.22M impressions, which at their $0.50 CPC was about $15K of ad spend they had not made.

Both numbers then roughly doubled over the following eight months, which is the part that matters. Paid traffic is linear: spend twice as much, get twice as much. Organic compounds, because the pages that ranked in month six are still ranking in month sixteen while new ones are being indexed on top of them.

Where it stands today

76.6K clicks and 2.2M impressions over the sixteen months, at a 3.5% average CTR and an average position of 9.6. The CTR is up from 2.6% at the checkpoint and the average position improved from 10.2 to 9.6 while the number of ranking pages was still growing, which is the combination you want: more pages, ranking better, converting harder.

On the money: 76,600 clicks at the $0.50 they used to pay Google is about $38K of traffic. We quote the saving as $30K deliberately, because not every organic click is one they would have bid on — some are research, some are queries no sane campaign would buy. Even at the conservative number, the traffic they did not pay for is several times what the work cost.

The number the client actually cares about is not in Search Console: 20 to 30 phone calls a day, inbound, from people who found them on Google. We are still working together and the site is still growing.

What we would tell you before you start

• This is not a fast channel. The curve above is flat for weeks before it is steep, and switching ads off on day one takes nerve. If you cannot survive a quiet quarter, taper instead of cutting.
• Programmatic SEO needs real data underneath it. Cevik Motors had genuine stock, per make and model. Generate pages without something real behind them and Google treats the whole site as scaled content, which is a hole you do not climb out of quickly.
• Volume without indexing is nothing. Publishing tens of thousands of pages is the easy half; getting them crawled and kept in the index is the half that decides whether any of it earns.
• Measure the business, not the dashboard. Clicks are a proxy. Calls are the result, and the only reason this case study exists.