How to evaluate a stock scanner: 5 questions before you pay.
Most scanner sales pages are 50 filters and a screenshot. Here are the five questions that actually predict whether a tool will be useful in six months — and how Finviz, Trade Ideas, Zacks, and Tapeline answer them.
I've signed up for almost every prosumer stock scanner since 2018. Most fail the same way: the tool is fine for a week, then you realise you have no way to tell whether the calls it surfaces are actually working. By month three you've added it to the pile of $20-$50/month subscriptions you keep forgetting to cancel.
If you're shopping for a scanner, these five questions will save you the cycle:
1. Can you see the methodology?
If the answer is "we use a proprietary blend of signals" you're being sold magic. The two questions you can't answer about magic are "is this working?" and "will this still work next month?" Tipranks, Zacks, Kavout, WallStreetZen all hide theirs. Tapeline names all six factors and shows each one's contribution per ticker on /how-it-works.
2. Where's the public scorecard?
Newsletter shops have known for 30 years that you should hide your losers. Mark Hulbert built a career being the only neutral grader of newsletter performance because everyone else hid the data. Look for a tool that auto-publishes every call it makes against the next-day market move — not a curated highlight reel. We do this at /scorecard; almost nobody else does.
3. What does the data come from?
"AI-powered signals" usually means "we bought a feed from a third-party market-data feed and slapped a score on top." Which is fine — that's also our spine. But know it. Bloomberg Terminal at $32k/yr uses similar feeds; the premium is the speed and breadth of their proprietary chat and curated news, not the raw data. Anyone charging $200/month for "exclusive AI signals" is reselling a third-party market-data feed and a third-party data feed.
4. Is the cheapest tier real?
Test it. If the free or cheapest paid tier strips out so many features the product is unusable, the team is incentivised to upgrade-trap rather than retain. Tapeline's answer is to put the whole published record outside the paywall entirely: the daily Top 10, every recorded top-10 pick with its next-session result vs SPY, a page per scored ticker, and the raw CSV/JSON — no account, no card. A free account is an email and a password, and it opens the scanner at the top ten scored rows. The card belongs to the 30-day Premium trial, which is a step you take on purpose ($0 that day, first charge on day 30, one click to cancel). Judge the product on the record before you decide whether to open an account.
5. Can you cancel in one click?
If you have to email support to cancel, that's a tell about how the team treats you generally. Stripe-portal cancel-in-one-click is table stakes; if it's not there, leave. (Yes, ours is.)
If a tool can't answer questions 1 and 2, walk away regardless of price. They're cheap to ask and predict 80% of the future regret.
See it in the scanner.
30-day Premium trial — starting it takes a card, $0 is charged that day, and one click cancels before the day-30 charge. Signing up is an email and a password, and lands you on the free plan. Or read the public record instead: the daily Top 10, the full scorecard and the raw CSV/JSON need no account at all. The scoring formula above runs on every scored US stock and ETF on each worker pass during US market hours, over prices delayed about 15 minutes.