The backend was working. The dashboard was working. The companion app was on a phone showing a banner. And the whole thing was called a placeholder that I had typed into a directory name months earlier and never revisited.
Naming turned out to be the hardest single-day problem in the project, and the marketing site that followed turned out to be the most opinionated design work I have done in years — because in this category, the default design language is something I actively wanted to reject.
Every good name is taken, and you find out too late
I had a name I liked. It was two syllables, it said what the product did, it sounded like a company. I was attached to it within about an hour, which is the dangerous part.
Then I checked the domains. The .com was registered. The .app was registered. Not parked-and-for-sale-at-a-reasonable-price registered — registered and in use.
The lesson, which every experienced founder already knows and I had to learn by losing a day: check availability before you fall in love, not after. A name you have already imagined on a landing page is a name you will rationalise keeping, and the rationalisations are all bad. getname.com. name.io. nameapp.com. Every one of those is a permanent small tax on every conversation: spelling it out loud, correcting people who typed the wrong thing, competing in search results with a company that owns the obvious domain.
So I started over with a rule: any candidate had to have both .com and .app free, and it had to survive being said out loud to someone who had never heard of the product.
Kidslen came from combining the two halves of what the product is: kids and lens. A lens is the right metaphor for what I had built — something you look through to see clearly, held deliberately, visible to everyone in the room. Not a camera hidden in a smoke detector. The name had to carry the transparency argument, because the transparency argument is the product.
kidslen.com and kidslen.app were both free. I bought both, for two years, through Cloudflare — which was already where the tunnel and DNS lived, so the whole public surface of the product sits with one provider I can reason about.
Two practical notes from the naming day:
Check the technical identifiers too. A brand name has to become an Android package name, an iOS bundle identifier, and a metrics prefix. Kidslen became app.kidslen for the package, app.kidslen.companion for the bundle, and kidslen_ for the Prometheus metric prefix. Those are extremely annoying to change later — package identity is tied to app-store listings and signing — so settle the name before the first signed build, not after.
Decide your market before your domain. I am marketing in English first, with Vietnamese as a second language, and I deliberately did not buy a country-specific domain. Buying domains for markets you are not entering is a recurring bill for a decision you have not made.
What every competitor’s website looks like
Before designing anything, I looked at the marketing sites of the established products in parental monitoring. There are a lot of them, and they are remarkably similar. The pattern:
- A friendly geometric sans-serif, usually rounded, usually the same handful of typefaces
- A palette of green or blue — green for “safe”, blue for “trustworthy”
- Illustrated children, flat-vector, smiling, in the hero
- Big trust numbers: millions of families, hundreds of thousands of downloads, a large figure with a plus sign after it
- Award badges and press logos in a row under the fold
- A hero headline about protecting your child, often with a threat implied just underneath
It is a coherent visual language and it is not stupid — it works, in the sense that it is what buyers in this category expect to see. But looking at fifteen of these sites in a row produces a specific feeling, which is that they all look like the same company. And the thing they most look like is a category, not a product.
More importantly, two of those conventions were closed to me for honest reasons. I do not have big trust numbers, because I have not got millions of families, and inventing them is exactly the kind of lie that a product about trust cannot tell. I do not have awards or press, because there are none. Every site in the category had those blocks, and I had nothing true to put in them.
That constraint turned out to be the most useful design brief I could have received.
Going the other way
If I could not compete on the category’s visual language, the right move was not to do a worse version of it. It was to look like something else entirely.
A soft serif instead of a geometric sans. The headings use a serif with warmth to it, paired with a humanist sans for body text that has full Vietnamese diacritic support — a non-negotiable, since half the content is Vietnamese and typefaces that handle stacked diacritics badly look broken to a Vietnamese reader. A serif reads as considered, editorial, adult. It signals that there is a person and an argument behind the page, which is precisely the pitch.
A warm cream palette instead of green or blue. Cream background, a teal that is muted rather than corporate, a coral accent. It looks like a well-made book rather than a security product. The category uses green to say “safe”; cream says “calm”, and calm is closer to what a parent actually wants to feel.
An original illustration instead of stock photography. No smiling stock family on a sofa. A drawn hero image, made for this product, that shows the actual concept — the visible, shared lens rather than the hidden eye.
Four honest trust pillars instead of invented statistics. Where competitors put “trusted by 3 million families”, the site puts four claims that are true and checkable:
| Pillar | What makes it verifiable |
|---|---|
| Always visible | The app shows a permanent banner and has no hidden mode — the child can confirm it |
| Minimal by design | The transparency screen lists exactly what is collected, and the list is short |
| Yours to take or delete | Export and delete are one click in the dashboard, not a support request |
| Consent on the record | Enrollment stores a versioned consent the family can re-read |
Each one is a claim the product either honours or visibly fails to honour. That is a very different kind of trust signal from a number nobody can check, and it is the only kind I am entitled to use.
I want to be clear that this is not pure virtue. Designing against the category is also differentiation, which is the only strategy available to a product that cannot outspend the incumbents. If you look like everyone else and you are smaller than everyone else, you lose. Looking different is a cheap way to be remembered — as long as the difference is real, which in this case it is: the design is different because the product is different.
The SEO an honest product still has to do
Being honest does not exempt you from being findable. The site is built with Astro, which means it ships static HTML with no client-side framework in the critical path, and that removes most of the hard part of performance work. The rest is discipline.
Per-page titles and descriptions, written by hand. Not templated from the H1. Every page — home, guide, FAQ, contact, each legal page — has a title that would make sense as a search result on its own and a description that is a sentence a human would read, under the length where engines truncate it. This is the highest-value, lowest-glamour SEO work there is, and most sites skip it because it is tedious rather than difficult.
Structured data. JSON-LD for the organisation, the software application, and the FAQ page. This does not make you rank; it makes what you already rank for render more usefully. A FAQ block that expands in the results answers a parent’s question before they click, and for a product whose main obstacle is suspicion, answering the question early is the point.
hreflang for two languages. The site is English and Vietnamese, with equivalent pages in both. Without hreflang annotations, search engines treat those as competing near-duplicates; with them, they understand that one is the localisation of the other and serve the right one. The annotations have to be reciprocal and complete — every page pointing at its sibling and at itself — or they are ignored, which is a failure mode that produces no error message anywhere.
Fast fonts. Two typefaces, self-served with the exact character subsets needed, font-display set so text renders immediately in a fallback rather than sitting invisible, and the primary face preloaded. Web fonts are the single most common cause of a slow-feeling page in an otherwise well-built site, and a serif-led design is exactly the kind that gets this wrong.
Real content. The guide section explains how the product works, with figures and a short video, in both languages. Not a blog full of keyword filler about screen time. The honest version of content marketing is writing the documentation a buyer needs before buying, which for a monitoring product is mostly “what exactly will my child see?” — and answering that well does more for conversion than for ranking.
The uncomfortable part
There is a tension I have not fully resolved, and I would rather name it than pretend.
A marketing site’s job is to make people want something. A product about monitoring children has an obligation not to make people want it too much — not to trade on parental fear, which is the most reliably effective lever in this entire category and the one I have watched competitors pull hardest. Every site that opens with what could happen to your child online is using fear as a conversion mechanism.
I decided the site does not do that. It does not describe dangers. It does not imply that a parent who has not installed something is failing. The hero talks about what the product is and what the child sees, and the strongest emotional claim on the page is about having a conversation rather than preventing a catastrophe.
This is almost certainly worse for conversion in the short run. I think it is the only version that is defensible, and I also think — though I cannot yet prove it with numbers, so I will not pretend to — that it is the version that produces families who stay, because they bought a tool rather than a reassurance.
Next in this series, and the one the whole thing has been building toward: how to actually sell it. Who pays for parental software, what they are really buying, and how a solo builder competes with companies that can outspend him a thousand to one.
The site is at kidslen.app — serif, cream, no invented numbers.