Invisible on the map, competing with yourself abroad

Two failures that look nothing alike

A company can lose search traffic in two completely unrelated ways, and the repairs have almost nothing in common. The first failure is geographic. Somebody standing four streets away types a query that ends in the words near me, and what comes back is a competitor whose map card is filled in, whose opening hours are current and whose phone number matches the one on their website. Nothing about your own site is broken. You simply were not in the register that the map was reading from.

The second failure is stranger, because it only happens to companies that have already done something ambitious. A site publishes the same page in three languages. All three versions get indexed. None of them carries an instruction telling the search engine which reader each version was built for. The result is that your own pages line up against each other for the same query, and the engine picks whichever one it guesses at, which is very often not the one that would have converted.

The VITON13 catalogue keeps these apart deliberately, because they are not the same work performed at different scales. Local SEO and cards in Yandex and Google Maps sits in the Local SEO category at $200 – $800 for a one-time setup. Multilingual SEO and hreflang settings sits in International SEO at $400 – $1 500. A third entry, Multilingual version of the site, is filed under Localization inside the web development group, because producing a second language version is construction work rather than search work.

Both failures are silent, and that is the whole difficulty with them. Neither produces an error page, a broken build or an angry email. The map you are missing from renders perfectly for everyone else. The duplicate language versions load, look correct and pass every technical check a developer would think to run. The only visible symptom is traffic that never arrives, which is the hardest symptom in the world to notice.

What a map card fixes that your website cannot

The catalogue describes the local entry plainly: the company is found by searching for near me and in maps, branch cards are filled out, reviews are processed, the data is the same everywhere, and for offline businesses this is a direct flow of calls. Read that description slowly and notice how little of it happens on your own domain. Card fields, review threads and registry data live in systems you do not control and cannot deploy to.

That is the whole reason the service exists as a separate line. A well-built site with a contact page, a schema block and an embedded map still does not put you into the map result, because the map result is drawn from a business listing, not from a crawl of your pages. The listing is a record with its own fields: name, category, address, hours, phone, photos. If nobody fills those fields, they stay empty, and an empty record loses to a complete one every time.

The phrase in the description that carries the most weight is the data is the same everywhere. Consistency is the part companies get wrong without noticing, usually through ordinary organisational drift. The office moves one floor up and the suite number changes on the website but not in the listing. A landline is replaced by a mobile in one place and not the other. Hours change for summer in one register and stay winter in the other. Each individual mismatch looks trivial and none of them is.

The catalogue rates this item Beginner, which is worth reading as a statement about the skill required rather than about the value returned. Nothing here demands architecture decisions or a migration plan. It demands somebody who will sit down with the registers, complete every field, reconcile the mismatches, work through the review threads and then check that the same string appears in all of them. It is unglamorous, finite work, which is precisely why it goes undone for years.

One location, three to five, and the money in between

The catalogue prices this job at 15 000 – 50 000 ₽, and it is the only entry of the three that publishes a breakdown inside the price itself: one location around 25 000, a network of 3–5 around 50 000. In dollars the same work is quoted as $200 – $800 for a one-time setup. The word one-time matters as much as the numbers do, because it tells you this is a project with an end, not a monthly retainer that renews until you cancel it.

Now do the arithmetic the published figures invite, as an illustration rather than a measurement. If a single location lands near 25 000 and a five-point network lands near 50 000, then four additional locations account for roughly 25 000 between them, which works out at something in the region of 6 000 apiece on that example. This is arithmetic performed on catalogue figures, not an observed outcome, not a quoted per-branch rate, and not a promise about what your own network will be quoted.

The same shape appears in the published time. The duration reads one day for a location, two to three days for a network of three to five points. Handled strictly one after another, five locations would consume five days. The catalogue says two to three. The reason is structural rather than heroic: the taxonomy, the category choices, the description copy, the photo set and the review policy are decided once and then applied to each card, so the second through fifth points inherit most of the thinking done for the first.

For a buyer, the practical consequence is that scale is the cheapest thing about this service and hesitation is the most expensive. A company with three branches that orders the first card now and the other two next year pays the setup thinking twice and gets the discount for neither. The catalogue publishes both the single-location figure and the network figure precisely so a reader can see, before any conversation happens, which side of that line their own business sits on.

Three language versions and no instruction about who gets which

The second entry describes its own failure mode more clearly than most sales copy manages: language versions stop competing with each other, each market sees its own version, duplicates do not eat up traffic, and the search engine correctly determines who to show what. Every clause there describes something being repaired, which means every clause describes something that was broken first. That is unusual honesty for a catalogue line, and it is the most useful sentence on the page for diagnosis.

Here is the mechanism underneath it. Without a declared relationship between versions, a search engine sees three pages that say approximately the same thing at three different addresses. It does not know they are siblings. It may treat two of them as near-duplicates of the third, it may consolidate signals onto whichever version happened to attract links first, and it may serve a reader in Madrid the English page because that page has the stronger history. Nothing has crashed. You are simply losing the argument with yourself.

The service is priced at $400 – $1 500, or 25 000 – 90 000 ₽, and the catalogue files it under International SEO with a duration of one to three days. The category placement is the tell. This is not translation and not front-end work. It is the job of declaring, page by page and market by market, which version belongs to which audience, and then verifying that the declarations are consistent in both directions so that the set actually resolves.

Notice what the description does not claim. It does not promise more traffic, a ranking position or a percentage. It promises that the versions stop competing and that the engine gets a correct answer to a question it was previously guessing at. That is a smaller and far more checkable claim, and it is the right one to buy on, because it describes a state of the site you can inspect rather than an outcome that depends on your competitors.

Building the version and wiring the signal are separate purchases

The third entry, Multilingual version of the site, is described as second and subsequent language versions with correct URL, hreflang and language switcher, so that the site begins to collect search traffic in other countries rather than showing Google translation to foreigners. It costs $300 – $900, or 15 000 – 40 000 ₽, takes 12-25 hours, is rated Experienced, and sits in the Localization category within the web development group rather than the SEO one.

The group assignment explains the relationship between this line and the previous one better than any diagram would. This is a build. Somebody creates the routes, decides the URL structure, wires the switcher so it lands on the equivalent page instead of dumping every visitor on the home page, and emits the hreflang markup as part of that build. It is finished when the second language version exists and behaves correctly, and 12-25 hours is a real build window rather than a configuration afternoon.

The International SEO line is what you buy when the versions already exist, usually because they accumulated over several years and several contractors. One was added for a trade fair, one because a partner asked, one because somebody translated the marketing pages and stopped. Nobody ever declared the relationships between them, or somebody declared half of them and the half-finished declaration is worse than none. That is an audit and reconciliation job across an estate, not a build, which is why the catalogue quotes it separately.

Read together, the two lines answer the two versions of the same question a buyer asks. If the other language does not exist yet, the relevant number is $300 – $900 and 12-25 hours. If it exists but the set is behaving badly, the relevant number is $400 – $1 500 and one to three days. A company that has already built three versions and then buys a fourth build has bought the wrong item for the problem it actually has.

Difficulty here is not measured in days

Put the three difficulty ratings next to the three durations and something counterintuitive falls out. Local SEO is rated Beginner and takes one day for a location. Multilingual version of the site is rated Experienced and takes 12-25 hours. Multilingual SEO and hreflang settings is rated Advanced, the hardest label of the three, and takes one to three days. None of the three runs longer than a few days, so the hardest label is plainly not a statement about the calendar.

Difficulty in this catalogue is clearly not a synonym for elapsed time. It is closer to a measure of how much damage a wrong decision does and how long that damage takes to surface. A badly filled map card is fixed by editing the card, and the fix takes effect quickly. A wrongly declared set of language relationships teaches search engines something false about your site, and you then wait for them to unlearn it, which happens on their schedule rather than yours.

That asymmetry should change how you procure, not just how you read. The Beginner item is safe to hand to whoever has the time and the patience, because the failure mode is a slow one and the correction is cheap. The Advanced item is the one where the difference between somebody who has wired a multi-market site before and somebody following a tutorial shows up months later, in traffic that quietly went to the wrong version and never came back to be counted.

It also explains why the highest price ceiling on the page sits on a job measured in single days. One to three days at $400 – $1 500 is not a rate you are paying for the hours. It is a rate you are paying for the decisions made during those hours: which version is canonical for which market, which pages are genuine equivalents, which pages exist in one language only and should therefore be declared as nothing at all.

What the three descriptions promise, and what they refuse to promise

None of the three catalogue entries contains a number about your results. There is no traffic multiple, no ranking guarantee, no percentage lift and no timeframe in which any of it is supposed to happen. Every number on the page is either a price or a duration, which is to say every number describes what VITON13 does rather than what the market subsequently does. That absence is a deliberate editorial position and it is worth naming out loud.

What the entries do promise is a set of verifiable states. Branch cards are filled out. Reviews are processed. The data is the same everywhere. Each market sees its own version. Duplicates do not eat up traffic. Second and subsequent versions have correct URL, hreflang and language switcher. Every one of those is a claim you can check yourself after delivery, without an analytics account, without waiting a quarter and without taking anybody's word for it.

The one outcome sentence in the set is carefully bounded: for offline businesses, this is a direct flow of calls. It states the mechanism by which map visibility turns into revenue for a business that answers a phone. It does not say how many calls. A supplier who tells you how many calls a card will produce, before seeing your city, your category and your competitors, is telling you something they cannot possibly know.

Which of the two failures you actually have

Start with a question about physical geography rather than about search. Does a customer ever need to know where you are in order to buy from you? If somebody drives to you, walks in, or receives a visit from you within a defined service area, then the map register is part of your distribution and its absence is a hole in your distribution. The catalogue floor of $200 for a one-time setup is small relative to a single recovered customer for most such businesses.

Then ask a question about language. Does your site exist in more than one language today? If the answer is no and you want customers in another country, the line you need is the build at $300 – $900 and 12-25 hours, and the hreflang work is part of what that build produces. If the answer is yes, and especially if the versions arrived at different times from different people, the line you need is the $400 – $1 500 audit and reconciliation.

Both answers can be yes, and for a company with branches in one country and customers in several others, both usually are. In that case sequence by reversibility rather than by price. The local work is finite, cheap to correct and produces its effect near-immediately once the register updates. The international work sets signals that take time to propagate and longer to unwind, so it is the one worth doing carefully rather than quickly, even though it is the shorter job of the two.

The failure worth avoiding above all is buying the international line to solve a local problem. A perfectly wired hreflang set does nothing whatsoever for a query that ends in near me, because that query is answered from a business register your markup never touches. Companies do this more often than you would expect, because international work sounds more strategic and the local work sounds like data entry. The register does not care how strategic anything sounds.

Four checks you can run before you spend anything

First, open your own site in a browser, view the page source of any page that exists in more than one language, and search it for the string hreflang. If nothing is found, the versions have never been declared to anybody. If something is found, check that the page it points at also points back, because a one-way declaration is treated as an unverified claim and ignored, and a half-wired set is exactly what the $400 – $1 500 line exists to reconcile.

Second, take the name, address and phone number exactly as they appear on your contact page and compare them character by character with what your map listings show. Not approximately, character by character. Abbreviations, suite numbers, the presence or absence of a country code and a different phone line for the branch are all the kind of drift that the phrase the data is the same everywhere is written to close.

Third, click your own language switcher from a deep page rather than from the home page. If it takes you to the home page in the other language instead of the equivalent page, you have a switcher that loses the reader's place, and that is a build defect belonging to the 12-25 hour Localization line rather than to the SEO one. It also tells search engines that the two versions are less related than they really are.

Fourth, count your locations and your language versions and write both numbers down before you talk to anyone. Those two counts are what move a quote inside its published band: one location around 25 000 against a network of 3–5 around 50 000, and a two-language site against a five-language estate. Arriving with the counts already established turns the first conversation into scoping rather than discovery, and shortens it considerably.

Reading the ranges before you ask for a quote

Every price on this page is a band, and each band has a shape you can reason about. The local line runs 15 000 – 50 000 ₽ and tells you exactly what moves you along it: the number of points. The international line runs 25 000 – 90 000 ₽ without publishing a driver, which is honest, because the driver there is not one countable thing but the interaction between the number of versions, the number of pages and how much of the set was previously wired wrong.

The build line is the tightest of the three at 15 000 – 40 000 ₽ and 12-25 hours, which is what you would expect from a job with a defined deliverable. Producing one additional language version of an existing site is a known quantity: routes, structure, switcher, markup, and the content that fills it. The variance between the low and the high end is mostly the size of the site being duplicated and how cleanly the existing build separates content from layout.

The most useful preparation, then, is not a brief describing what you want to achieve. It is a short factual sheet: how many physical locations, how many language versions today, how many exist only in one language, how many total pages, and which registers already have some version of your listing. Every one of those facts moves a quote inside its band, and every one of them is something you can establish in an afternoon without any supplier's help.