This page explains how a toy gets evaluated on Smart Toy Choice, what our labels mean, and what we refuse to do. It is the page we point to when a manufacturer disagrees with us.
The short version
We are a research desk, not a test lab. We do not claim to have played with everything we write about, and we never pretend otherwise.
When a product is genuinely in our hands, we say so at the top of the article, along with who used it, for how long, and with a child in which age band. Everywhere else, our conclusions rest on sources you can open yourself — and we link them, so you can check our work instead of trusting it.
That is a lower bar than hands-on testing. It is also a bar we can actually clear, which is the point.
What we check, on every article
Six things, in this order. Most pages that recommend toys check the first one and stop.
- The specification against the paperwork. Not the marketing page — the manual PDF, the accessory list, the certification filing. Manufacturers routinely put “batteries not included”, “app required” or “subscription sold separately” in a footnote, and the headline price never mentions it.
- The safety record. Recall notices from the US Consumer Product Safety Commission and the EU Safety Gate, plus the specific injury descriptions buried in owner reviews. A toy can be compliant and still be the reason a four-year-old needed stitches.
- Whether the product still exists as a service. Last app update, whether the company is still trading, whether the subscription is still on sale. A connected toy without a working backend is a plastic brick with a speaker in it. This is why our first article was about a robot that no longer works.
- The age claim against real use. What the box says, against what parents report about the age at which a child can actually use it without an adult sitting next to them. These two numbers are frequently two years apart.
- The real total cost. Sticker price, subscription, consumables, accessories, and how often the thing is discounted — because a product permanently marked down from $199 to $129 has a real price of $129, and the higher figure is theatre.
- What owners say at three months and beyond. Failure rates, customer service response, parts availability. The complaints that appear in month one are usually about setup. The ones that matter appear in month four.
How we label a verdict
We do not publish star ratings or out-of-five scores. A number implies we measured something, and in most cases we did not. Instead every article ends in one of four labels:
| Label | What earns it |
|---|---|
| Recommended | All six checks pass and we found no blocking problem. We would spend our own money on it. |
| Recommended with conditions | Worth buying, but only with a stated condition attached — an adult present for the first hour, a subscription you have to budget for, a specific age band, or a region where support exists. |
| Not recommended | We found a blocking problem: a dead service, a false age claim, a safety pattern, or a total cost that changes the value calculation. The reason is stated in the first paragraph, not buried. |
| Insufficient evidence | We could not verify enough to say either way. We publish this label rather than guess. It appears more often than you would expect. |
We publish “not recommended” and “insufficient evidence” often. A site where nothing is ever bad is not checking anything.
When we do have the product in hand
Some articles are hands-on, and those are labelled as such at the top of the page with a line that states who used the product, for how long, and which age band the child was in. A hands-on line is a factual claim, not a badge, and we only print it when it is true.
Manufacturer samples are disclosed in the same place. A sample never changes a verdict. We do not accept a sample in exchange for coverage, and receiving one does not oblige us to publish anything.
What we will not do
- We will not write as though we tested something we did not. No invented anecdotes, no “after three weeks our seven-year-old…” on a product we never unboxed.
- We will not paraphrase a press release and call it a review. If the only available information is the manufacturer’s own copy, we say so.
- We will not recommend a product we could not verify. Where the evidence runs out, the label is “insufficient evidence” and the article says why.
- We will not let a brand see copy before publication, and no brand can pay for a verdict, a ranking, or a place in a list.
- We will not publish generated imagery as if it were a photograph of the product. Where an illustration appears, it is credited as an illustration.
How we handle being wrong
When we get something wrong we fix it and date the correction. Substantive corrections — a wrong specification, a safety-relevant error, a verdict that changed because the product changed — are noted in the article itself rather than quietly edited. A toy that needs a subscription, or an app that stops updating, becomes a different product after ninety days, and articles are revisited when that happens.
If you own a product we have written about and your experience contradicts ours, that is useful. Send it through the contact page.
Commercial arrangements
Our articles contain affiliate links and we earn a commission on some purchases. That funds the research. It does not affect labels, rankings, or which products we choose to cover. The detail is on the affiliate disclosure page.