SEO case studies
Automatic transmission repair, professional cosmetics, freight transport, retreats, a cosmetology clinic, a shop selling printing consumables. All on Yandex, Moscow and the regions, rank tracking on a live keyword set. The numbers come from Metrica, Pixel Tools and the client’s end-to-end analytics, each tied to a domain and a date.
How the work itself is done is on the SEO service page. This page shows only what came out of it.
Methodology
The case studies below are compiled from the SEO team’s working reports: Yandex Metrica screenshots for specific counters and Pixel Tools ranking exports with the tracking dates shown. Nothing has been recalculated or rounded in our favour.
Two case studies — jkclinic.ru and gevarius-shop.ru — come not from rank tracking but from Metrica reports and the clinic’s end-to-end analytics. What exactly was taken and from where is stated in the case study itself.
The label on a case study and on a card means we built the website ourselves first and then took it on for promotion. These domains also appear in the work portfolio.
The most expensive and most crowded niche in Moscow car servicing. The query “automatic transmission repair” gets 3,250 impressions a month, with 778 thousand documents in the search results. Two client websites sit in the top five at the same time.
The project grew out of a 2023 story: an automatic transmission website was brought up from zero, and reaching the top took five months. After that the work focused on holding positions and expanding the keyword set — from short commercial queries to the long tail of makes and components: CVT, robotised gearbox, torque converter, DSG, specific models.
Quarterly Metrica snapshot from 20 August to 19 November 2024: 9,033 visits and 8,053 visitors, 28.8 thousand pageviews. From search — 6,875 visits, three quarters of all the website’s traffic. There were practically no ads in this period: the “ad traffic” channel shows one visit.
Pixel Tools measurements for Moscow, snapshots from 1 October to 1 November 2024. Project visibility is 77.7%, up +69.8 points over the period. 264 queries grew, one dropped.
A neighbouring project in the same niche shows the same pattern from a different start: visibility 59.8% with growth of +57.8, nineteen queries in the TOP-3, one hundred and forty-nine in the TOP-10. The key commercial phrases — “automatic transmission repair price”, “urgent automatic transmission repair”, “automatic transmission repair Moscow” — went from 100+ into the top ten in a month.
In the results for “automatic transmission repair”, fourth and fifth place are held by two different clients of ours — akpp.studio and akpp-balance.ru. For a service owner this means both spots in the niche are taken by someone other than a competitor.
A topic agencies usually call a dead end: seasonal demand, fuzzy keywords, search results crowded with aggregators and booking platforms.
The keyword set is about nine hundred queries: vipassana, retreat near Moscow, yoga retreat, retreat centre, pranava master, branded phrases. Pixel Tools tracking across two search engines, measurements from 1 September to 8 November 2024.
At the start of tracking almost nothing was in the TOP-100. By November: 158 queries in the TOP-3, 549 in the TOP-10, 759 in the TOP-20, 848 in the TOP-100. Over the whole period two queries dropped. Visibility grew from near zero to 39.1%.
In Metrica for 2 September – 12 November the search channel brings 1,845 visits. In a niche where a deal is a paid tour rather than a parts order, volume matters less than the fact that the visit comes from a query with the intent to go.
krasota3.ru is a professional cosmetics shop with a range of several dozen brands. Before our team took over, the website had been run by another agency for a year with no growth.
Quarterly Metrica snapshot from 20 August to 19 November 2024: 61.4 thousand visits from search, 14.3 thousand direct visits, 1,948 internal referrals. 64.2 thousand new visitors, 8,145 returning.
The work focused on the shop’s branded keyword set: mesaltera, storyderm, q lab, skinasil, DMAE, lipolytics, Jessner peel. Tracking across two search engines, measurements from 1 September to 1 November 2024: twenty-one queries in the TOP-3, one hundred and fifty-seven in the TOP-10, three hundred and fourteen in the TOP-20. 319 queries grew.
The search results for the Natinuel brand are a separate story. There the shop ranks fourth, above marketplaces and niche aggregators.
natinuel-cosmetics.ru was launched as a separate single-brand shop for the Italian brand Natinuel — in search results where the manufacturer’s official website already ranks.
The website was built on an SEO structure from the start: keywords were collected before the layout, and sections and landing pages were designed around the keyword set rather than adjusted after launch.
The quarterly Metrica chart by week shows a start from zero in early October 2024 and growth to 729 search visits by mid-November: 713 new visitors, 123 returning. Six months later, in the week of 12–18 May 2025, the website already brings 563 visits a week, with 88.5% new visitors.
In the search results for “Natinuel cosmetics” the website is in second place, right below the official site of the brand’s Russian distributor.
A B2B project where decisions take a long time and a lead is expensive. Weekly Metrica measurement from 1 January to 19 May.
Over the period: 68 thousand visits, 61.1 thousand visitors, 82.6 thousand pageviews. From search — 49.2 thousand visits, direct visits 12.1 thousand, paid traffic 4,084 visits.
The weekly curve rises without dips: from about a thousand visits a week in January to over four thousand by May. Over the same time the paid channel shrinks — search replaces ads rather than adding to them.
The only case study in this selection where the “before” point was recorded publicly and in advance: on 1 April 2025 the team published a rankings snapshot and promised to show the result exactly one month later.
Pixel Tools snapshot as of 1 April 2025: visibility 0%, zero queries in the TOP-3, TOP-10 and TOP-20, ten in the TOP-50, six beyond 100. The keyword set includes “Nespresso milk frother”, “original Nespresso coffee” and similar.
Three weeks later, in the 21–27 April snapshot: 348 visits and 294 visitors in a week, +48.7% on the previous week. Page depth 6.3 pages, time on site 4 minutes 24 seconds, bounce rate 10.6% — rare numbers for an online shop.
The team’s honest comment on this project from 5 May: the first month brought no sales, the keyword set turned out to be too competitive, and reaching the top will take another one or two months. We show the case study exactly as it is.
jkclinic.ru came to us with an unfinished migration to new URLs: the same pages lived at two addresses at once, canonical tags pointed to other sections, and some links returned 404. The only case study in the selection where we counted leads, sales and revenue rather than visits.
The previous contractor started moving the website structure to new URLs and never finished. What remained was a broken structure: the same pages existed at two addresses at once, canonical tags pointed to other sections, some links returned 404, and there was no commercial markup at all. For a search engine this is a signal of distrust — it sees a contradictory website and ranks it lower.
We completed the migration and removed duplicates across the two URLs. Removed the wrong canonical from 10 pages. Converted 404s into 301s while keeping ad tracking parameters. Restored the sitemap: 221 → 231 pages in the index. Then commercial factors — markup for the clinic, services, prices and FAQ, rewritten titles and snippets on 37 pages, booking blocks and internal links to priority services, and a doctors feed for Yandex Medicine covering 16 specialists.
June 2026, the first full month of work. Organic search brought 34 leads: 24 qualified, 22 booked, 18 converted to a sale. Revenue 261,710 RUB, average order value 14,539 RUB, lead-to-sale conversion 53% — higher than most of the clinic’s other channels. Organic traffic for the same month year on year: 2,126 → 2,495 visits, +17%. Non-branded CTR in search 2.09% → 2.46%.
The leads, sales and revenue here come not from our counter but from the clinic’s Roistat end-to-end analytics. We don’t publish a snapshot of that report: it shows the clinic’s entire revenue across all channels, not just the search row. Of our own tools, only Metrica is shown under the case study.
This is the first full month of work, not a year’s result: in June 2026 organic search brought the clinic 18 sales and 261,710 RUB. Next comes the season, and that will be a different conversation.
gevarius-shop.ru sells fasteners and consumables for printing, binding and branding: screw posts and screws, ring binding mechanisms, piccolo rings, covers, cords, ribbons and elastic bands with tips. Promotion across all of Russia, Yandex and Google, June 2023 – December 2025.
In summer 2023, when work started, the shop had practically no organic traffic. Search visitors started coming in the autumn of that year. Over the whole period of work, June 2023 – December 2025, search brought 145,504 visits and 51,531 visitors, with a bounce rate of 11.72% and page depth of 2.99.
Traffic comes from both search engines at once: Yandex — 96,209 visits (66.12%), Google — 48,097 (33.06%). Baidu, Bing and Rambler together account for less than one percent. Geography covers all of Russia: visitors came from 960+ cities, Moscow accounts for about 46%, followed by St. Petersburg, Nizhny Novgorod, Yekaterinburg, Rostov-on-Don, Krasnodar and Novosibirsk.
What we did: a technical audit and error fixing — speed, indexing, response codes, robots.txt and sitemap. Catalogue structure and landing pages matched to demand, human-readable URLs. Keyword collection and clustering, meta tags and headings for categories and product cards. Internal linking, filling and updating content, commercial factors. A separate line of work was SEO blog articles for informational queries; one of them is the first organic result for “piccolo rings”.
For three category queries the shop is the first organic result on Yandex; in two cases only an ad block labelled “promo” is above it. There was no keyword-set rank tracking in this project — the case study shows live search results, not an export from a tracking system.
Approach
What all the projects above have in common is a sequence of work, not a set of one-off fixes.
Technical first, content second
Until a website passes on speed, indexing and structure, adding more text is pointless. On projects where the CMS can’t cope, we say so before the contract, not after three months of work.
Keywords before layout
On new websites the keyword set is collected before the mock-up, and the section structure is designed around demand. That’s how the single-brand shop that reached traffic within a month was built.
Rank tracking from day one
The keyword set is put on regular rank tracking for the right region straight away. Without it, a quarter later you can’t tell growth from seasonality, and the conversation with the client comes down to feelings.
Behaviour matters more than volume
Bounce rate, depth and time on site are viewed together with rankings. Traffic that doesn’t read the page through won’t hold rankings in 2025.
A weekly report for every project
The same set of metrics, the same comparison week. The report arrives regardless of whether the week was good or bad.
Turning down projects we can’t deliver on
We don’t take on promotion for websites on custom-built CMSs of doubtful quality, or for sites that would be easier to rebuild. Saying no is cheaper than burning through the budget.
Honestly
We deliberately don’t fill in what we didn’t measure. Here is what was left out.
Money and leads — except in one case study
Revenue and sales were counted in exactly one project — jkclinic.ru — and not by us: they are tracked by the clinic’s own end-to-end analytics. We don’t publish a snapshot of that report because it shows the clinic’s entire revenue across all channels. For the other eight projects we have no revenue figure of our own: for two of them the clients said the website paid for itself in the first month, and the case study says “according to the client”, not an amount. In most projects the Metrica goals are set up as a formality — form submission and phone click — and conversions in them are close to zero. The difference between jkclinic.ru and the rest is not that we worked better there, but that it has end-to-end analytics.
“Before” points for older projects
Ranking growth is counted from the first measurement in the tracking system. If the keyword set was put on tracking a month after work started, part of the growth didn’t make it into the number. The only project with a publicly recorded “before” point is the coffee shop, where a zero-visibility snapshot was published before work began. A separate case is gevarius-shop.ru: the annual snapshots there are real, but together they don’t reconcile with the report for the whole period, and we say so directly in the case study instead of making one fit the other.
Failures and niche limits
In the coffee shop the first month brought no sales — the keyword set turned out to be too competitive, and the team said so directly. In some young projects weekly traffic is still measured in tens of visits. That is a normal working stage, but we won’t promise quick results at it.
If your project needs revenue and lead figures, not just traffic, that is settled at the start: we set up goals, connect forms and calls, and a quarter later the conversation is already about cost per lead.
Portfolio
For every project the same set of Metrica metrics is pulled once a week. Below is a snapshot of two spring 2025 weeks for the projects run in parallel.
Below are 13 of them, the ones run in parallel in this snapshot.
The label on a case study and on a card means we built the website ourselves first and then took it on for promotion. These domains also appear in the work portfolio.
Some of the projects in the table are young websites in their first months, with tens of visits a week and three-digit growth. That is a normal stage, and we show it as it is rather than hiding it behind portfolio averages.
Next
We crawl the pages, track rankings, compare you with the top results and assess each section. Free, on an hour-long call. If the job isn’t a fit for us, we’ll say so right away.
Write to us