How to evaluate a competitor price monitoring tool
Evaluate a competitor price monitoring tool by asking for proof, not features: can you see the source page and timestamp behind every price, how often does it really read, how are products matched and what happens to uncertain matches, are currencies kept separate, which guardrails exist, can it change prices without you, and does it keep a record of its own mistakes.
The nine questions
Ask each vendor the same list and write the answers down. A vendor who cannot answer one in a sentence has told you something. Melqart’s answers follow each question so you can hold us to the same standard.
- Can I open the source page and see the timestamp behind any price you show me? — Melqart: yes; every observation carries URL, UTC timestamp, content hash and an archived snapshot. Observations taken before receipts existed are labelled as such rather than back-filled.
- How often do you actually read each competitor, and can I see the next scheduled read? — Melqart: weekly, daily or six-hourly by plan; next sync time is shown per source, and fleet-wide source health is public at /status.
- How do you match their product to mine, and what happens when you are not sure? — Melqart: brand, unit, currency and variant checks; anything not certified safe for pricing goes to a review queue and never into a recommendation.
- Do you convert currencies to compare? — Melqart: no. Like is compared with like; a cross-currency match is marked unsafe for pricing.
- Which guardrails run before a recommendation reaches me? — Melqart: minimum margin (when a cost is on file), price floor, maximum drop, daily velocity, implausibility; blocked drafts stay visible with their reason.
- Can the tool change my prices without me? — Melqart: no. Every write is approval-gated, and in the current early-access release the write itself has not launched; approvals are recorded with evidence.
- What does it need from my store, and can I take my data out? — Melqart: read access through the platform’s official API with credentials you generate; a full export (products, observations, decisions, raw event log) is one click in Settings.
- What do you show me when you were wrong? — Melqart: the ledger keeps every recommendation and its measured outcome, including the ones that did not work. Outcomes are labelled practice-mode until real writes begin.
- What can I verify before paying? — Melqart: a free audit of your own public storefront, a live workspace on real data, public source health and a public market report — no card, no install.
What should you not be impressed by?
Detection-latency claims without a source-health page to back them; logos without a link to the merchant; “AI-powered” without an explanation of what the model decides and what a person decides; and any promise that prices will change while you sleep. Melqart has removed claims of its own that failed that test, and lists the changes on the changelog.
Questions people ask next
How long should a trial be?
Long enough to see two full read cycles at the cadence you would pay for, so you can check a recommendation against its evidence at least once. On a weekly plan that is two to three weeks.
Is Melqart in beta?
Yes — early access. What is live, what is practice-mode and what is coming is stated on every relevant screen, on the status page and on the changelog.