How we get a number, and what we print when we can't get one
KidTally publishes one thing: what kids phones and their plans actually cost, with the date we last checked. That only means anything if you can see how the checking works — so this page is the whole procedure, including the parts where it fails.
There is no bylined author here, and that is deliberate rather than an oversight. Nobody on this site is presented as a person who has tested every device, because that would not be true. What stands behind the figures is the procedure below, the dated source link on every one of them, and a public log of every change we have ever made.
How a figure gets on the site
Every price traces to the vendor's own pricing page, linked from the device page it appears on. We do not take prices from retailers, from other review sites, or from a vendor's marketing email, because none of those is the thing a parent will actually be charged.
A script refetches those pages every Monday at 06:00 UTC. For each device it extracts prices two independent ways: the structured data the brand publishes on its own page, and a parser written against that page's layout. Both have to agree. When they disagree, nothing is published — a disagreement usually means the brand redesigned its page, and a parser reading a redesigned page confidently returns the wrong number.
Three more checks stand between the script and the site:
- A handset has to fall between $30 and $2,000, a plan between $3 and $80 a month, an activation fee between $0 and $150. Anything outside those bounds is a parsing error wearing a dollar sign.
- A change of more than 40% from the last recorded value stops the pipeline. Real prices move, but they rarely halve overnight, and when they do it is usually a promotion with an end date rather than the price.
-
Whatever fails a check is marked
needs-reviewand waits for a person. Reviewing only the exceptions is the reason the review still happens in week forty.
A figure that clears all of it carries the Verified stamp and the date it cleared. The stamp is a claim about one specific thing — that on that date, that number appeared on that vendor's page. It is not a claim that the price is still true today, which is why the date is printed next to it rather than hidden in a tooltip.
Why some cells say “—”
A dash means we could not find that figure published anywhere. Right now 1 of the 5 phones we track is missing one of the two numbers that decide the cost, and the reason is structural rather than secretive: a bring-your-own-carrier handset has no plan price of its own, because the plan is whatever your carrier charges you.
A dash is a claim about us as much as about the vendor, and on 2026-08-10 it was the wrong claim twice. We had recorded Gabb's plan price and Troomi's handset price as unpublished, and said so on the site, when both were on the vendor's own site all along: Gabb's in its help centre, Troomi's on its shop subdomain. Our pipeline checked one page per vendor and we reported what it saw as what existed. Both figures are now in the tables, both corrections are in the log below, and the data file records two source URLs per vendor instead of one.
We do not fill those gaps with an estimate, a "starting from", or a figure copied out of a competitor's review. The temptation is real, because a table with a hole in it looks unfinished next to a table full of confident numbers. But an estimate presented in the same typeface as a verified price is indistinguishable from one, and the entire value of this site is that you can tell the difference at a glance.
Where a price does exist but not where you would look for it, we print the figure and say which page it came from, because "on the vendor's site somewhere" is not the same as findable. Which brands put prices where, and what that does to anyone trying to compare them, is set out in price transparency.
When we get one wrong
Every change to every figure is written to a public log: what changed, the value before, the value after, the date, the source URL, and whether a machine or a person confirmed it. Nothing is edited silently, and corrections are not quietly folded into a "last updated" date. The most recent entry is dated 2026-09-07. You can read the whole log, and download the underlying data as CSV or JSON, on the price tracker.
When something we published turns out to have been wrong rather than merely out of date, the correction goes into the same log, in the same format, with our previous figure still visible in the "old value" column.
Who pays for this
Nobody, currently. There are no affiliate links on KidTally today and no paid placements. No vendor has seen a page before it went up, and none has any say over what a page says.
That will change. Affiliate links are the intended way to fund the work, and when they appear they will be marked and disclosed on the pages carrying them. The order of operations is what stays fixed: a figure gets on the site by passing the checks above, and a commission is not one of the checks. Every figure that has ever moved is in the public log with the date and the source, which is what makes this checkable instead of merely promised.
If you think a number here is wrong
Tell us, and bring the link. A figure on a vendor's live pricing page beats anything on this site, and we will take the correction and log it. Vendors are welcome to do the same; the only thing we will not do is remove an accurate figure because it is unflattering.
Write to hello@kidtally.com.
What we deliberately don't publish
- No scores or star ratings. Nobody here has rated these devices, so a number out of ten would be the same fabrication as an invented price.
- No stock or availability claims, because we do not check inventory.
- No “best for” verdict we cannot support. Where we compare devices, the comparison runs on the published figures and says which one costs less over a stated period. Which phone is right for your child is a different question, and we try not to let the first answer stand in for the second.
Start with the first phone decision guide, see what each device lets you monitor in the monitoring matrix, or work out a total with the cost calculator.