See What Your Magento Store Ranks For, in Plain English

Start with the disclosure, because it changes how you should read everything below. EveryHost built Sondavo. It is our product, we make money from it, and this article exists partly to tell Magento merchants it is there. We are not reviewing it from the outside and we are not pretending to be neutral about it.

Here is why it exists. Several times a month a hosting customer would arrive with a version of the same question: traffic is down, or a rival is above us, or the store feels invisible, so is it the server? Sometimes it genuinely was, and we could prove that in minutes. Far more often it was not, and the honest answer was that we had no idea, because we could see response times and error rates and nothing at all about search. So we built the thing we kept wishing we could hand people: tell a site owner where they stand, what is moving and what is broken, using their own data wherever possible, in words a shop owner can read without a glossary. Judge what follows on whether the checks are useful to you, not on our enthusiasm for our own work.

The name is nautical. A sounding measures how deep the water is under a vessel, and Sondavo asks the equivalent question of a website: how deep does this site sit in search? Position one is the surface, everything below it is depth, and the further down a page sits the less light reaches it. Position eleven is not one worse than position ten, it is a different world, because it is page two.

The free check at sondavo.com needs no sign-up and returns three things. Where you stand: the searches your site already appears for, plotted at their true depth rather than flattened into an average. What is rising, which the product calls shoaling, borrowing the word for a seabed coming up toward the surface. Those are the keywords sitting just under page one, ranked by a deterministic engine so every score traces back to the rule that produced it. And what is holding you back: a polite crawler surveys the site the way search engines read it, and reports what it finds in plain English.

For a Magento merchant the middle one is usually what to act on this month. A keyword at position thirteen is one decent internal link, one honest category description and one fixed canonical tag away from page one. A keyword at position ninety is a project. Knowing which terms are shoaling turns a vague ambition to do some SEO into a short list of specific pages.

The most expensive SEO mistake in Magento is not a bad decision, it is an absent record. A store moves to a new server, or upgrades from 2.4.6 to a current release, and four weeks later somebody says traffic is down. Down from what? Nobody wrote it down, so the conversation that follows is guesswork, and fixes that come out of guesswork are expensive.

Both events can genuinely move rankings, for mechanical reasons. An upgrade can change URL keys, regenerate the sitemap, alter how canonical tags are emitted on filtered pages, and shift the theme's markup enough that Google recrawls the lot. Our guide to the Magento 2.4.9 hosting requirements covers the stack side of that work. A migration adds its own hazards: old category addresses remapped once, then remapped again during a tidy-up, leaving two-hop and three-hop chains that dilute the signal at every step. Our Magento migration service handles the redirect map as part of the move for that reason, and the redirect checker is how you verify afterwards that the map does what the plan said.

So take the baseline before the work starts. Record the searches you appear for and at what depth, save what your top twenty pages earn in clicks, and run the canonical, robots, sitemap and header checks on one product page, one category page and the homepage. It costs an hour. Afterwards, when somebody says traffic is down, you are comparing against evidence instead of memory.

A growing share of the questions that used to start on Google now start in an assistant, and plenty of them are shopping questions: best waterproof work boots for cold storage, who supplies replacement seals for this pump, which UK retailer stocks this brand in a size fourteen. The assistant answers with a shortlist of names. You are on it or you are not, and unlike a search result there is no page two to be on.

Checking this properly is harder than it sounds, because asking once proves nothing: assistants vary their answers between runs, and location changes them too. Sondavo's AI visibility feature puts your market's real questions to four engines, ChatGPT, Claude, Gemini and Perplexity, plus Google AI Overviews, on a schedule. It stores the answers verbatim so you read the actual words rather than a score, shows who was named, and lists the pages that were cited. It runs UK-located where the engines support that, which matters for a British retailer whose rivals in a US-located answer are firms that do not ship here.

The second half is more of a build and hosting question. Sondavo also scans whether AI agents can use your site at all: whether the menus can be opened, whether the links can be read, whether there is an llms.txt summary, and whether robots.txt takes a deliberate position toward AI crawlers rather than an accidental one. A heavily JavaScript-driven Magento storefront can be perfectly usable by a human and close to opaque to an agent, and a robots.txt written in 2019 may be blocking assistants you would like to be cited by. For the traffic side there is a related feature that reads your own server log and separates the AI agents that can cite you from the ones only collecting training data.

This is the part we are most willing to be judged on, because it is where the category most often disappoints people. The rule is stated plainly: no invented numbers, ever. Every figure comes from one of three places. Your own Google data, read through Google's official read-only Search Console API, so clicks and impressions are exact rather than modelled. Your own site, crawled by Sondavo's own crawler. Or a licensed third-party index, kept separate from your own data and always labelled as an estimate.

Two consequences follow. When a data source is not connected the interface says so, rather than filling the gap with a plausible chart. And there are no simulated previews or sample data dressed up as your results, a discipline that costs some polish on day one and is why the numbers mean anything on day thirty. Anything you check stays private by default.

The opportunities list works the same way. Fixed rules, not a language model, so the same data in produces the same list out, and every item carries its evidence and the rule it fired on. If it says a page has a canonical problem, it shows you the tag. That matters when you hand work to a developer or an agency, because a finding you can point at is one somebody can act on without a debate about where it came from.

The eight tools above are free and unconditional: no sign-up, no card, no email. Beyond them a free plan at zero pounds covers one site and, in the product's own words, is not a trial: it works forever. That is a weekly site survey, three months of history and three AI answers a day, still with no card.

The paid plans buy frequency, depth and breadth rather than access to hidden results. Starter at nineteen pounds a month surveys every three days, holds six months of history, allows ten AI answers a day, and adds CSV export and email alerts. Professional at forty-nine pounds a month is where most single-store merchants land: ten sites, a daily survey, thirteen months of history, competitor tracking, AI visibility, twenty-five AI answers a day, three teammates and API access. Agency at one hundred and forty-nine pounds covers fifty sites, twenty-four months of history, fifteen teammates and white-label client reports.

The wider product adds up to sixteen months of your own Search Console history, a site survey reporting eighteen kinds of technical fault in plain English, a page-by-page ledger putting each finding next to the clicks that page earns, competitor standings, verified backlinks and long tail clustering. Site fixes are separate, quoted first, with no fee if a fix does not pass a re-check.

Sondavo is not a Magento product. It works on any website and knows nothing about Magento's internals. It will tell you a category page carries a canonical tag pointing somewhere unexpected. It will not tell you which admin setting, extension or theme override produced that tag, because it never sees inside your store. Everything it reports comes from outside, the way a search engine sees you.

It also does not replace a developer. Magento's SEO surface, URL rewrites, layered navigation config, sitemap generation and store-view scoping, is genuinely intricate. What a good diagnosis does is let you brief a developer precisely instead of paying them two days to find out what is wrong. It turns a vague worry into a specific ticket.

And it will not fix a slow store. No amount of canonical tidying compensates for a category page that takes four seconds to respond. If the checks come back clean and the store still feels invisible, look at speed next.

Some of what search engines and AI agents judge you on is squarely a hosting matter. Response time under crawl load is one: a store that answers quickly for a human browsing three pages can crawl badly when Googlebot requests two hundred URLs a minute, and a crawler that times out indexes less of your catalogue. Correct status codes are another, and so is serving a redirect map cleanly rather than through rewrites bolted on over the years.

On EveryHost plans the stack is built to your store's required versions rather than one fixed recipe, and sized against your real catalogue and traffic. If you are moving a store, our managed migration includes the redirect map and the post-move verification, the single most common place rankings get lost. More broadly, our Magento 2 hosting UK plans cover the server, the stack and the go-live, and an engineer will talk the plan through before you commit to anything.

Either way, run the free checks first. They cost nothing, they take minutes, and they will tell you whether the problem you are about to spend money on is the one you actually have.