Brand watch — typosquat & lookalike monitoring
Brand watch generates lookalike variations of each domain — typos, homoglyphs, TLD swaps — and checks which ones actually resolve in DNS. When a new lookalike goes live, it alerts you exactly once. It's a separate panel and alert, not a health signal: it never changes the 0–100 grade.
For each domain, brand watch generates around 40 lookalike candidates using a deterministic, hand-rolled set of rules: homoglyph and character swaps (0↔o, 1↔l, rn→m, vv→w), QWERTY-adjacent typos, dropped characters, inserted characters, transposed characters, and TLD swaps (.com↔.net, .org, .co, .io, .info, .app, .online, .site, .xyz). It operates on the second-level label and understands multi-part suffixes like co.uk and com.au.
It then DNS-resolves each candidate — with bounded concurrency, a 4-second timeout, and fail-open behavior so a slow lookup never blocks the run. A candidate counts as live when it resolves to at least one A or AAAA record. Every live lookalike is persisted as its own row, and Domain Watchdog alerts exactly once per newly-discovered live lookalike. The whole thing runs inside the daily health scan.
Brand watch is not a health signal. It never changes the domain's 0–100 score or A–F grade — it's a separate panel with its own alerts. A flagged lookalike is information about impersonation risk, not a mark against your domain's own health.
It only flags lookalikes that are actually live in DNS — registered and resolving — not every theoretical permutation. That keeps the noise down: you hear about a lookalike someone has actually stood up, not the thousands of variations nobody has registered.
Coverage is capped at about 40 candidates per domain, so brand watch is a strong early-warning net, not an exhaustive trademark sweep. It catches the common, high-probability impersonation patterns; it isn't a substitute for a full brand-protection service if you need one.
Frequently asked questions
How does brand watch find lookalike domains?
It generates about 40 lookalike candidates per domain from deterministic rules — homoglyph and character swaps, QWERTY typos, omissions, insertions, transpositions, and TLD swaps — operating on the second-level label (and handling suffixes like co.uk). Then it DNS-resolves each one to see which are actually live.
Does a lookalike domain affect my health grade?
No. Brand watch is deliberately separate from the health grade — it never changes the 0–100 score or the A–F letter. It's its own panel and its own alert, because an impersonating domain is a risk to your brand, not a defect in your domain's health.
Will it flag every possible typo of my domain?
No — it only flags lookalikes that actually resolve in DNS (registered and live), and it caps candidates at about 40 per domain. That means you hear about lookalikes someone has actually stood up, not every theoretical permutation, and it's an early-warning net rather than an exhaustive trademark sweep.
How often will I be alerted about the same lookalike?
Once. Domain Watchdog alerts exactly one time per newly-discovered live lookalike, and the alert routes through the same engine as your other alerts. You won't get repeated notifications about a lookalike you already know about.