How we count a “search” (methodology deep-dive)
Seekmodo’s billable-search definition — typeahead vs SERP, bot filtering, and why the methodology is public.
Billing fights start when “search” is undefined. One vendor counts keystrokes. Another counts API calls. A third counts “sessions.” Finance asks for a forecast; engineering shrugs; the invoice still arrives.
This post explains how Seekmodo defines a billable search in plain language. The canonical, footnoted reference remains /docs/methodology — use that page when you need source anchors for comparative advertising. Here we focus on intuition and edge cases merchants actually argue about.
The unit we sell
Seekmodo meters shopper search submissions against published monthly tiers (Hobby 5,000 → Scale 5,000,000), after bot checks. A submission is a completed search request that hits the gateway as a search — not every character typed into a typeahead box, and not every admin click in the console.
That is why our compare pages contrast Seekmodo with Doofinder-style per-request metering (keystrokes, assistant turns, indexer writes). Same human behavior; different billable atoms. See the worked example in Doofinder pricing vs search-volume billing.
Typeahead vs full search
Typeahead / suggest UIs often fire many lightweight requests while the shopper is still typing. Seekmodo’s product and metering design treat the search submission as the unit finance can explain — aligned with “someone searched for X,” not “someone hovered on the sixth letter of X.”
Implementation details evolve; the methodology page is authoritative when you need the precise rule set for a contract conversation. The principle is stable: we optimize for an invoice that matches merchant intuition.
What should never bill like a shopper
Bots, scrapers, and abusive clients are filtered by the gateway’s bot checks where possible. Load tests and monitoring should use documented non-billable or sandbox patterns — not production shopper quotas. If you are load-testing, talk to us or use a sandbox tenant so you do not burn a customer’s monthly allotment.
Admin merchandising actions (pinning, synonym edits) are not “searches” in the shopper sense. MCP tool calls have their own plan allowances on the published ladder — see /pricing feature notes.
Multipliers you should know
Some product surfaces consume more than one search-equivalent unit (for example image search and certain chat/assistant tool calls on lower tiers). Those multipliers are published on the pricing feature matrix so Hobby buyers are not surprised. Text search, suggestions, and recommendations stay at 1× on every tier in the published copy.
Overage and Enhanced Native
When a tenant exceeds its monthly search allotment after trial, Seekmodo’s connector story is fallback to Enhanced Native rather than a blank search box. That is an operational definition of “what happens when you run out,” not only a pricing footnote. Read what Enhanced Native means for the free-on-install half of that story.
How to forecast
Use the pricing calculator with last month’s search volume from your dashboard (or a conservative estimate during trial). Prefer one tier with headroom over constant overage anxiety. For comparative math against keystroke vendors, keep methodology footnotes open beside the calculator.
If you are writing an RFP answer, cite /docs/methodology and the relevant /compare page rather than paraphrasing this blog post from memory.