• Commercial SEO · Screaming Frog

Tool Shop SEO: Features Matter More Than Adjectives

The promotion of the tool online store starts with the route of the search bot, not with the texts: first it should reach from the root address to the product card, then the power, voltage and type of the cartridge should lie in the card - in text, not in the image. In the 6,486-card ProTool store we measured, the certificate is valid and the domain root returns a 301 to insecure HTTP. The site map there is generated in 34.2 seconds and contains each product an average of 3.9 times. As long as this is the case, the descriptions work at full capacity - no matter how many of them you write.

See how it works
Price
after a free audit
Guarantee
30 days after project sign-off
Measured niche directory
6,486 product cards and 361 categories, measurement 07/31/2026
after a free audit
Price
30
Guarantee

days after project sign-off

6,486
Measured niche directory

product cards and 361 categories, measurement 07/31/2026

25,537
Records in the site map

on 6,857 unique addresses — 3.9 copies of one product

34.2
Site map generation

seconds on the fly, the address is non-standard

100%
The language of addresses on the map

Russian-speaking with Ukrainian by default

in the two measured niche stores of no Product block
Product marking
free store inspection, 3-5 business days
Before the estimate
What's included

Complete list of work and what you get as a result

  • We go through the route from the root address to the product card and fix each redirect and response code - then this list is used as an emergency, before everything else.
  • We reduce the site map to one source with the correct address in robots.txt: the product is included once, and not under each path of the category.
  • We match the language branch we are indexing with the one given by default - the signals stop contradicting each other.
  • We collect semantics around the parameters of the catalog: power, voltage, type of power supply, configuration, compatible equipment, and not around adjectives.
  • We install landing pods for "instrument type + parameter" connections with their own addresses and headers - only for those that have a demand for data.
  • We close the rest of the filter combinations with canonicals and bypass rules, so that instead of benefit, we do not get thousands of almost identical pages.
  • We transfer the characteristics to the text table of the card - it is from these fields that the filter, marking, and model comparison are then collected.
  • We put the Product markup with price, availability, brand and article number and BreadcrumbList on the way to the card: we describe only what is actually on the page.
  • We measure the speed of the card separately from the main card and keep the card itself in operation - search traffic comes to it, and the main card is usually optimized.
  • Every month we give a report with an unchanged set of numbers: what was done, the dynamics of crawling and indexing, the plan for the month.
When this service isn't right

What's not included — so there are no surprises at delivery

  • Transferring characteristics from manufacturers' catalogs to your database — working with data is counted separately.
  • Shooting and processing photo tool.
  • Contextual advertising, price aggregators, marketplaces.
  • Promises of items by brand names: both the manufacturer and the networks compete with you for them.
  • Pricing and assortment: The site does not make the expensive position competitive.
Who it's for

Situations where this service delivers results

Scenario 1 of 5

There are thousands of positions, and there are hundreds in search

The goods are filled, the categories are laid out, and the tour goes in a circle. In the measured professional tool store, the sitemap yielded 25,537 records for 6,857 addresses — each item is listed under all category paths, an average of 3.9 times. Cleaning does not add traffic, it stops spraying it.

We'll review your situation in a free audit

Free inspection of the tool shop

We go through the store the way a search bot goes: from the root address to the product card. Losses most often lie in this route - not in the texts, but in the fact that the bot sees a different version of the site than the buyer. Access is not needed for this: everything is visible from the outside, from what your server gives.

What we measure

  • Route from the root addressWhere does a bare domain lead and by what protocol. The real find: the certificate is valid, but the root returns a 301 to insecure HTTP - and anyone who manually typed the domain gets there.
  • Site map and time of its generationAt which address it is located, how many records are in it and how many unique addresses. One of the measured stores returned 25,537 records for 6,857 unique URLs, and the map itself took 34.2 seconds to build.
  • The language version is in the indexWhat language does the search engine see? We saw a store where 100% of the addresses on the site map were in Russian, although the site displays Ukrainian by default.
  • Characteristics as textAre the power, voltage, revolutions per minute, type of cartridge included in the text table of the card - or are they only on the image from the manufacturer.
  • Filters and landing pagesDoes the popular "instrument type + parameter" relationship have its own address, or does it only exist as a filter state. And aren't all the combinations open at once by chance?
  • The speed of the card is separate from the main oneThe main one is often optimized, but the card is not. We measure both and show the difference: in the measured niche store, the first response is 671.2 ms with 407.8 KB of main markup.
  • Marking and countersWhat blocks of structured data actually are. In the two niche stores that we measured, there is a description of the organization and search on the site, but there is no description of the product.

What you get

  • Root-to-card bot route with all redirects and response codes.
  • Directory numbers: entries in the site map, unique addresses, duplication rate.
  • A list of what's breaking indexing right now, in order of priority.
  • A 30-minute conversation on the document, where you can argue with the conclusions.

Timeline: 3-5 working days

Why is it free

Because half of the finds in tool stores are broken technical trifles, not a lack of content. It would be impolite to sell texts where the root of the site resets to HTTP.

What's next

After checking, we divide the work into two parts: emergency (what is losing traffic right now) and planned. We take the emergency ones into operation first and separately — we don't have to wait for the approval of the entire scope for them.

Short form: your contact and site URL

  • Contract, act and 30-day warranty

    Every project gets a written contract: scope, deadlines, amount, acceptance procedure. After delivery — act and invoice, then 30 calendar days of warranty.

  • Sole proprietor & bank transfer

    The contractor is a registered sole proprietor. Payment by invoice with closing documents.

  • Rights & access — yours

    Code, design and materials transfer to you after full payment. Domain, hosting, repository and analytics are registered to you.

  • Client portal instead of email chains

    During the project you get access to a portal: contracts, invoices, acts and project status in one place.

  • European clients

    Among our work — projects for Norway, Bulgaria, Moldova and Spain.

  • Verifiable numbers

    Every case in the portfolio comes with a link to a live site and a technical measurement.

  • Audit first, then pricing

    There is no price list on the site intentionally: the scope of the same work differs multiples between clients.

  • We say "no" when unsure

    If the task isn't ours or the deadline is unrealistic — we tell you upfront.

What affects the price

Why two seemingly identical tasks are priced differently

  • How many items actuallyWe count the catalog by two methods — by the site map and cross pagination of the standard search — because these numbers differ. In the measured store, the card declared 25,537 entries, and the live product pages turned out to be 6,486. The volume sets the second number.
  • The state of the technical part at the startHere is the main fork of the estimate. A store where redirects are the goal and the site map is one, immediately enters the semantics. A shop that builds a map in 34 seconds on the fly and duplicates goods four times over requires separate engineering work first, which the shop next door might not have at all.
  • Where do the specs come from?Filled fields in the database are transferred to the card template. Parameters that exist only as a picture in the manufacturer's catalog have to be disassembled by hand, and this is a separate line, not part of SEO.
  • The number of passengers according to the parametersEach such page has its own semantics, its own title, its own set of products and its own link. Ten connections and one hundred connections differ in many ways, so we first narrow the list down to those that are in demand.
  • Language versionsThe second branch duplicates URLs, meta tags and lines in the sitemap. And it is not closed by the translation of the interface: in order for the language to work for sales, there must be local payment and delivery nearby.
  • Access to the serverPart of the emergency findings is fixed by the configuration, not by the admin. If there is a contractor with a weekly queue between us and the server, this extends the deadline exactly by his queue.
Cases

Tasks and results in numbers — all metrics measured by us

online store of professional tools, catalog of over 6,000 items

Task
Understand why a large catalog is poorly indexed, despite a full set of payments and working analytics.
Solution
Traversal from the root address to the map, analysis of the site map with the calculation of unique addresses, verification of language branches and structured data.
Result
Measurement 07/31/2026: 6,486 product pages and 361 categories — with 25,537 entries in the site map at 6,857 unique addresses. The map is located at a non-standard address declared in robots.txt, it is generated in 34.2 seconds, each product gets into it on average 3.9 times under different category paths. All 100% of the addresses in it are in Russian, although the site defaults to Ukrainian. The first response is 671.2 ms with 407.8 KB of main markup. Separately, a critical issue was found: with a valid certificate, the domain root returns a 301 to insecure HTTP. From the structured data, there are OnlineStore, ContactPoint, PostalAddress, GeoCoordinates and SearchAction — the product is not described by any block.

B2B supplier of professional equipment for chocolate production

Task
Check how a small technical product catalog is organized in two language versions.
Solution
Parsing the sitemap index, checking the payment and delivery page, measuring speed, inventorying markings and counters.
Result
Measurement 07/31/2026: sitemap-index of four files — 106 products and 9 content pages in each of two language versions, 230 addresses together, no differences between languages. First server response 351.4ms, full markup download 490.4ms at 297.6KB. The six payment methods on the terms page are the broadest set in the measured sample. From the markup, only WebSite, SearchAction and ImageObject are present: product pages are not described by structured data. Calls are tracked by call tracking, but it is impossible to count them externally - the number remains with the owner.

Did not find your case?

Describe how it works on your side — we will tell you whether “SEO for the shop tool” fits and what it means in your situation. No brief and no call: one question, one answer.

Process steps

Transparent stages with approval at every step

Total duration:25–110 days

  1. Walkthrough and emergency finds

    3-7 working days

    We go through the route from the root address to the card: protocol, redirects, response codes, availability for the bot. What is losing traffic just now, we take out a separate list and start with it, without waiting for the approval of the rest of the volume.

  2. Site map and duplicate addresses

    3-10 working days

    We find out at which address it is located, how many records it contains and how many unique URLs there are. We transfer the generation to the rule "the product enters the map once" and compare the result with the live catalog. The number of pages in the index after cleaning may drop for a short distance - this is an expected trend, and we warn about it in advance.

  3. Language branch and canonicalization

    3-12 working days

    We fix which version we are indexing and reduce the site map, canonical addresses and internal links to it. A product that is available from multiple categories stops sharing weight between its own copies.

  4. Semantics for the parameters and structure of landings

    5-25 working days

    We take the parameters from your catalog and demand, not from the ceiling: power, voltage, cartridge type, tool size. We compile a list of connections that will receive their own addresses, and a separate list of what will remain as the state of the filter without an address.

  5. Product card: characteristics and marking

    7-35 business days

    We export the parameters to a text table, put Product with price, availability, brand and article number, BreadcrumbList on the way to the card. We describe only what is actually on the page: marking nonexistent is a risk of manual sanctions, not an advantage.

  6. Measurements, report and next cycle

    4-21 business days

    Re-circulation and the table "was - became" according to the same indicators: errors on the route, unique addresses on the map, the share of cards with a table of characteristics, the first answer on the card. How many pages are actually in the index is visible only in your Search Console - so you need access to it from day one.

Technologies & integrations

What we build on and what it connects to

Stack

  • Screaming Frog — crawling the directory with response codes and redirect chains; on tens of thousands of addresses it goes in parts
  • Bypassing the site map with the calculation of unique addresses - it is he who gives the duplication rate; but the number of records says nothing about indexing
  • Google Search Console is the only source where the number of indexed pages is visible; the report hides some of the queries
  • JSON-LD Product / Offer / BreadcrumbList — a chance for an extended snippet; Google decides to show it
  • rel=canonical — reduces card copies to one address; on filter combinations only works with bypass rules
  • nginx and .htaccess — things that cannot be fixed from the admin can be fixed here: protocol, redirects from the root, response codes
  • OpenCart — the sitemap in the basic build is built on the fly and pulls the product under each category path
  • GA4 - shows the organic trend, not the accounting truth

Integrations

  • Google Search Console
  • Bing Webmaster Tools
  • GA4
  • Google Merchant Center
Sodkova under the parameter against the state of the filter

How this option differs from the alternative

What the search engine seesa separate address with a title, a selection of products and text for a specific parameter
How many such pages to makeexactly as much as the connection gains demand in the directory data
The price of erroran empty page for a fictitious request, which is visible in the report and is removed
What you need to get startedthe characteristics fields in the database are filled out — without them, there is nothing to compile a selection from
When a filter is enoughwhen a request with a parameter actually exists and is repeated
What we need from you

We can't start without this — best to prepare in advance

  1. Access to Search Console and, if you have one, to Bing Webmaster Tools. The real number of indexed pages is visible only there, it cannot be counted from the outside.
  2. Access to a server or person who can change the configuration of redirects. Part of the emergency findings in this niche cannot be fixed from the admin.
  3. Decision, which language version is the main one. It's a business decision, and the entire address structure depends on it — it's expensive to redo it later.
  4. Technical characteristics in a form suitable for import or access to the supplier of this data: name, article number, parameters in separate fields.
  5. One person from your side with the right to say "yes" to the structure of the directory - otherwise every step will wait for a meeting.

If something is missing — let us know, we'll help you gather it or do it as a separate task.

FAQ

Most frequently asked questions — with concrete answers

Where to start if the catalog is already large?

From the bot route, not from the texts. In the Pro Tools store we measured, 6,486 products had a valid certificate and the root of the site returned a 301 to insecure HTTP — meaning anyone who typed the domain manually landed on the HTTP version. Until this is fixed, any texts work at half capacity. And separately: the certificate itself does not guarantee anything - it encrypts the channel, but does not make the site workable.

Why does our site map take a minute to generate?

Usually because it's built on the fly and includes every product under every category path. Our measurement on the tool store: generation 34.2 seconds, 25,537 records for 6,857 unique addresses, each item 3.9 times on average. This is treated not by a faster server, but by a rule according to which the product enters the map once. Just don't expect an increase from this: cleaning doesn't add traffic, it stops spraying it.

Do we need separate pages for "screwdriver 18v" and the like?

Yes, if such a request exists, and no - if you invented it yourself. We take parameters from the data of your catalog and demand, not from the ceiling, and make addresses only for those connections that are actually dialed. The rest remains the state of the filter without its own address. If you open all the combinations in a row, you will get thousands of almost identical pages - this is a deterioration, not an improvement, and you will have to rake it with the same canonical.

We have the characteristics in the picture from the manufacturer. is this a problem

For the buyer - no, for the search engine - yes. The text from the image is not readable, so the card, where all the details are in png, looks empty for search. It's the same story with the names of the carriers: in the measured garden tool store, the logos of the delivery services were pictures without a caption, so neither the search engine nor the screen reader could see them. Transferring parameters to database fields is a separate work with data, and we calculate it separately, and do not hide it in "SEO".

We have two languages. which to index?

The one they buy from you. The main thing is that the solution should be one and agreed upon: in the measured store, the site displayed Ukrainian by default, and all 100% of the addresses in the site map were in Russian. In such a situation, the search engine receives a contradictory signal and chooses itself, not as desired. And remember that a second language is an ongoing expense: doubled meta tags and texts that need to be kept alive, not a one-time job.

Do you work with stores of official sellers of brands?

Yes, this is most of our measurements in this niche. An important nuance: the presence of official status does not give an advantage in the search in itself - it gives the right to the assortment and to normal data about the product. The winner in the search is the one whose data is available in text and is not blurred by dozens of addresses. We do not promise items by brand name: both the manufacturer itself and large chains compete with you for them.

And will the Product markup raise us in output?

No, and no one will honestly say otherwise. The markup doesn't move the position - it gives a chance for an extended snippet with price and availability, and whether to show it is up to Google and can change its mind tomorrow. We do it for another reason: in the two measured stores, the niche structured data describes the organization, contacts and search on the site, and the product is not a single block. And a separate warning: describing with markup what is not on the page is a risk of manual sanctions, so we do not put reviews or ratings "for beauty" there.

How long does the first stage last and when will the movement be visible?

Emergency technical findings are closed in 1-3 weeks, depending on whether there is access to the server. Next comes the semantics and landing — it's 25-110 working days for the entire volume, and the term depends on the size of the catalog and on how many solutions need to be agreed on your part. Regarding "visible": the number of pages in the index after cleaning duplicates may initially drop, and this is normal. You need to look at it in your Search Console, access to which we need from day one.

Send the address of the store - we will go through it the way a search bot goes.

In response, a root-to-map route with response codes, your sitemap numbers with a duplicate ratio, and a list of what's breaking indexing right now. Access is not required for this.

From measured casesMeasurement 07/31/2026: 6,486 product pages and 361 categories

View cases
  • Reply within 2 hours
  • No commitment
  • We work under a contract

There is no price list on the site on purpose: the same work differs several times over between two clients, and a “from” figure explains nothing in that case. First a free audit — we count your pages, duplicates and speed — then we name the sum and the deadline and fix both in the contract.