reviewspytrend.com SpyTrend's claims, checked and dated

Language: EN RU

How we verify SpyTrend's claims

We take one public statement about SpyTrend at a time, find the place in the product where that statement is either true or false, measure it with the page state written down, and record a verdict with the date of the check. Four verdicts exist: confirmed, partly confirmed, contradicted, not verifiable. Nothing here is scored out of five.

Document
METHOD
Revision

Why a log instead of a review?

A review tells you what the writer concluded. A verification log tells you what was checked, where, when, and what came back — so you can disagree with the conclusion and still use the evidence. That is the whole difference in format between this site and the usual ad-spy write-up, and it is why every entry in our register carries an identifier, a date and a named method rather than a paragraph of opinion.

The unit of work here is a single statement, not a product. The vendor's page saying that coverage stops at Meta is one such statement; a third-party write-up saying that the price list is unpublished is another. Each gets its own record, its own verdict and its own timestamp, because they can be true and false one without the other, and because they age at different speeds. As it happens, the register contradicts both of those.

What counts as a claim?

A claim, for our purposes, is a statement about SpyTrend that is specific enough to be wrong. "The service is convenient" is not a claim — there is no observation that could refute it. "The geo filter contains 193 countries" is a claim, because someone can open the filter and count.

We take claims in the form the vendor actually publishes them, not in a form that is easier to defend. If a marketing page says coverage is limited to Meta platforms and TikTok is on the roadmap, that is the claim we check — we do not quietly soften it into "coverage is focused on Meta" so that it survives contact with the product.

We also check claims a vendor makes about other companies. SpyTrend's own comparison table quotes competitors' prices and search speeds. A price on a competitor's public page is checkable; a search-speed figure measured by the vendor on its own bench is not something we can reproduce, and it is marked accordingly rather than repeated.

Where do the claims and the evidence come from?

Three sources, captured on 4 August 2026
SourceWhat it gives usHow it was captured
Vendor marketing pages on spytrend.comThe claims themselves: the home page, the pricing page, the Connect to AI page, the company page and four blog posts about individual sectionsAutomated crawl over the sitemap, 29 pages, plus manual extraction of clean text from the Russian locale
The live product, signed in on a Pro accountThe counter-evidence: sidebar, dashboard counters, filters, the plans screen, and a section-by-section walkthroughManual walkthrough with a screenshot taken for each observation
Public profiles of third partiesContext for claims the vendor makes about other products — rating profiles parsed from each page's own data payload rather than read off the rendered pageMachine parse of profile data, 495 reviews across the profiles involved

Every entry in the register names which of these three it rests on. An entry resting only on the vendor's own page never earns the verdict confirmed — at best it is recorded as published by the vendor and left at that.

What does one check actually consist of?

  1. Quote the claim atomically. One sentence, in the vendor's own words, from a named page.
  2. Decide in advance what would refute it. If we cannot name such an observation, we say so instead of inventing a verdict.
  3. Go to the place in the product where the observation lives — a specific section, a specific filter, a specific screen — and record the state of that page: which source tab is selected, which filters are applied, whether the list is unfiltered.
  4. Measure. Count list items programmatically where counting by eye would be unreliable; read counters off the header of a clean page rather than off a filtered view.
  5. Capture a screenshot as the artefact of the observation, so a later reader can see what we saw rather than take our word for it.
  6. Write the verdict, the date and the page state into the entry. A number without the page state it was taken from is not a result, and we do not publish one.

What counts as confirmation?

Confirmation means we reproduced the thing being claimed, in the product, on a date we name. Seeing the same sentence repeated on a second marketing page is not confirmation — it is the same claim twice. This distinction does much of the work on this site: several statements about SpyTrend appear on two or three vendor pages and are still confirmed by nothing except each other.

For counted things — countries in a filter, sections in a menu, categories in a list — confirmation means an actual count, performed on the live interface, with the counting method stated. For prices, confirmation means the figure appears both on the public pricing page and on the in-product plans screen, because a price that differed between the two would be a finding in itself.

One thing is deliberately not treated as proof: our own memory of an earlier check. When an entry is revisited it is re-measured and re-dated, never carried forward on the assumption that nothing moved.

What do the four verdicts mean?

The verdict scale used on every record in the register
VerdictMeaningTypical trigger
ConfirmedWe reproduced the claim in the live product or on a named vendor page, on a stated date.The published price list matches the plans screen inside the account.
Partly confirmedTrue with a qualification the original wording omits — it holds in some sections but not others, or only in a particular page state.A filter mode present in two sections and absent from two others.
ContradictedThe live product returns the opposite of what the published statement says.A page describing coverage as Meta-only while the account carries a second source tab holding tens of millions of ads.
Not verifiableThe claim cannot be settled with our access: internal processes, infrastructure, private roadmaps, or measurements taken on the vendor's own bench.A stated sub-second search response, or a competitor speed comparison run by the vendor.

There is no fifth verdict for "we didn't like it". Judgements of taste belong in a review, and this is not one.

What do you do when the marketing page and the product disagree?

We believe the product and record the disagreement as the finding. This rule is not a preference — it follows from what each artefact is. A marketing page is a description written at some point in the past and updated when someone remembers to; the running product is the thing itself. When they diverge, the interesting fact is not which one is prettier but that the vendor's own description has fallen behind what it sells.

The largest divergence we found on 4 August 2026 has exactly this shape. The public page still describes coverage as Meta-only with TikTok on the roadmap, while the signed-in account carries a Meta/TikTok source switch in three sections and the TikTok tab returns 34,032,263 ads and 1,912,533 creatives with live statuses and dates running up to 08/03/2026. The claim is contradicted, and the direction matters: the vendor is understating its own product rather than overstating it.

The same rule cuts the other way when the product turns out thinner than the page promises, and we apply it identically. There is no softer verdict for divergences that happen to flatter the vendor.

Why does every entry carry a date?

Because unverified facts decay, and some decay within a single day. While collecting third-party rating data on 4 August 2026 we hit a clean example: a cached copy of one competitor's public profile carried a platform sanction — the rating suppressed for a guidelines breach, with a note about removed fake reviews — while the live version of that same profile on the same day showed the sanction lifted and the rating restored. Two honest observers reading the same URL hours apart would have published opposite facts. A dated observation survives that; an undated one is just an assertion.

Product counters decay the same way, only faster. The dashboard we read on 4 August 2026 reported 1,059,139,605 total ads and 4,360,481 new ones that day. Any of those figures quoted without its date misleads by construction, and quoted without its page state it is worse than misleading — two of our own counters came back different on a second pass purely because a source tab and a filter set differed between reads.

So the date is not decoration on the entry, it is part of the assertion. We never claim "this is true". We claim "this was true here, on this date, in this state of this page".

Why is there no rating out of five?

A single score compresses dozens of separate observations into one number, and the compression is where the information goes. Knowing that a service scored somewhere in the high fours tells you nothing about whether its price list is public, whether its geo filter behaves the same in every section, or whether its own marketing describes the product it currently sells.

Scores also invite a kind of laundering we would rather avoid: once a number exists it gets quoted without the checks that produced it, and it keeps getting quoted long after those checks went stale. An entry with a verdict and a date resists that — quote it and you carry the date along.

If you want a summary judgement, read the contradicted entries first. They are the shortest honest answer to the question of where this vendor's description of itself fails to match the thing it sells.

How often is this revisited?

Our rule is a 90-day horizon. An entry older than 90 days is either re-measured and re-dated, or visibly marked as stale — it does not silently keep standing as current. For the entries collected on 4 August 2026 that horizon falls on 2 November 2026.

Two categories are re-checked sooner than the horizon whenever we notice movement: anything about pricing, and anything where we recorded a divergence between a vendor page and the product. A divergence is temporary by nature — someone eventually updates the page — and an entry that goes on describing a divergence after it has been fixed is as wrong as the original claim was.

Every re-check appears in the revision history with its date, including re-checks that changed nothing. A revisit that confirmed the previous result is information too.

What is your conflict of interest?

This site is dedicated to a single product. That is a structural bias, and we would rather state it plainly than bury it: a site existing to discuss one vendor has an obvious interest in that vendor remaining worth discussing, and you should read everything here with that in mind. What we offer against it is not a claim of neutrality but the method above — dated observations you can go and repeat yourself.

The direct answer: there is one, and it is financial. This is an affiliate publication. Links to the vendor's site are affiliate links: following one may earn this desk a commission under the SpyTrend programme, which pays 10% of each payment made by a referred customer. We take part in it. Every such link carries rel="sponsored" and the campaign tag utm_source=revspytrend, applied by the builder rather than by whoever wrote the sentence — disclosure cannot fall off through forgetfulness.

What matters next is what that conflict can spoil and what it cannot. It can spoil selection: which claims we take into the register, which we quietly leave out, and how fast we refresh a record that has become inconvenient. The method does not fully protect against that, and it is fairer to say so than to pretend otherwise. What the conflict cannot do is move a verdict that sits on published evidence: the observation is named, the source is linked, the date is stamped, and any reader can repeat the check in ten minutes and see the discrepancy if one appears.

A testable consequence instead of a promise: of the twenty-two records, three contradict the vendor's public statements, six confirm them only in part, and five are marked unverifiable because the vendor publishes nothing to check them against. Fourteen of the twenty-two — the majority — are unhelpful to the vendor. A site where the commission outweighed the method would look precisely the other way round, and that difference is visible from outside, without access to our accounts.

The checks were performed on a paid Pro account bought with our own money: no free access, promotional tier or evaluation account was provided by the vendor, and the vendor did not know in advance what we intended to check.

We do not accept payment for a verdict, and no verdict on this site was shown to the vendor before publication. If that ever changes, it will be stated on this page with the date it changed rather than quietly omitted.

How do I get something corrected?

Write to [email protected] with the identifier of the entry — every record has one, and its permalink has the form #claim-C-01 — and what you believe is wrong with it. Evidence helps most when it is a screenshot with a visible page state, or a URL we can open ourselves.

Corrections are not applied silently. When an entry changes, the old verdict and the new one both appear in the revision history with dates, so anyone who quoted the earlier version can see what became of it. We would rather show a visible correction than a clean history.

This applies equally to the vendor. If SpyTrend can show that an entry is wrong, it is corrected on the same terms as anyone else's correction, and dated the same way.

Questions about the method

Do you test the product yourself, or only read the vendor's site?
Both, and the distinction is the point. Claims come from the vendor's public pages; evidence comes from a signed-in Pro account of the live product, walked through section by section on 4 August 2026 with a screenshot per observation. A claim supported only by the vendor's own page is never marked confirmed.
Why do you trust the product over the vendor's own marketing page?
Because a marketing page is a description written in the past and updated when someone gets to it, while the running product is the thing being described. When they disagree we record the product's behaviour as the finding and the disagreement itself as the story — including when the divergence flatters the vendor, as the largest one we found does.
What does "partly confirmed" mean in practice?
It means the statement is true with a qualification its own wording leaves out. The usual shape is a feature that exists in some sections of the product but not others, or a counter that holds only in one particular state of the page. The entry always names the qualification rather than leaving the reader to guess it.
Are you a certified fact-checking organisation?
No. We are an editorial desk applying a published method, not an accredited body, and we deliberately avoid the structured-data markup reserved for certified fact-checkers. What we offer instead is that every observation names its place, its page state and its date, so you can repeat the check without trusting us.
How current is this register?
The observations were made on 4 August 2026 and published on 10 August 2026. Our rule is a 90-day horizon, which for these entries falls on 2 November 2026: after that each entry is either re-measured and re-dated or visibly marked stale. Pricing changes and fixed divergences pull a re-check forward.