Stock Screener Buyer's Guide: Filters, Data and Price

“Which stock screener has the best data?” There is no single answer to that, because the best data for a dividend investor filtering on payout history is a different thing from the best data for a trader who wants relative volume refreshed during the session, and most tools are stronger on one side than the other. So start with the screens you actually run, and judge each tool by whether it runs those correctly.

What the money buys

Most screeners come as a free tier plus one or more paid tiers. The free version usually handles common filters: market cap, sector, P/E, yield, price and volume. Paid tiers tend to add depth. More fundamental fields, longer history, custom formulas, more saved screens, alerts, export and sometimes backtesting.

Whether the upgrade pays comes down to one question. Do you keep hitting a wall? If your screens run fine on the basic fields, the paid features are mostly convenience, and convenience is a fair thing to pay for as long as you know that is what you are buying. If you keep reaching for a filter the free version lacks, say several years of free cash flow margin, that filter is the thing to test before you pay.

Filters: count the ones you would use

Feature pages like to quote how many filters a tool has. Ignore the total. What matters is whether the handful of fields you rely on exist, are defined clearly, and can be combined the way you think.

Fundamental depth

Check whether you can filter on annual and quarterly figures separately, on trailing twelve months, and on change over time, such as revenue growth across three or five years. Look for cash flow and return measures like free cash flow and return on equity. Some tools stop at headline valuation ratios. That is thin.

Technical filters

Price against moving averages, relative volume, new highs and lows, percent change over a window you choose. Find out whether these are computed on daily closes or intraday. The same filter gives a different list at noon than it did at the prior close.

Custom formulas

The better screeners let you type your own expression, for example price to sales divided by the sector’s median price to sales, or a blend of two ranks. If you think in ratios the vendor never built a menu item for, this feature saves hours of spreadsheet work, and it is often reserved for paid plans.

Data is where tools really differ

Two screeners with identical menus can return different lists for the same screen on the same morning. The reasons are ordinary. Screener numbers disagree when one tool uses trailing earnings and another uses forward estimates, when one strips out one-time items, or when one refreshed last night and the other last week. Four things are worth checking.

• Coverage. Which exchanges are in the universe? Are OTC stocks, ADRs and ETFs included, and can you exclude them? A screen that quietly mixes OTC names with listed stocks can fill your results with shares that barely trade.
• Update frequency. Prices may be real time, delayed or end of day. Fundamentals may update the day a filing lands, or days afterward. The tool should say which, for each field.
• History depth. Multi-year growth filters and averages need years of clean data, so ask how many years come with the plan you would actually buy.
• Definitions. A glossary should give the formula, the period and any adjustment for each field. Without one, “P/E” on that tool could mean several things.

Output: after the list appears

A list of forty tickers is where the work starts. You need to sort by any column, including custom ones. Export to CSV matters if you track candidates in a spreadsheet. Saved screens let you rerun the same criteria each week and see what changed, and alerts tell you when a stock enters or leaves a screen without you having to check.

Backtesting deserves suspicion. Ask whether the test uses only the figures that were known on each past date, and whether companies that were later delisted or acquired stay in the historical universe, because dropping them makes almost any screen look better than it would have performed. Check output features on the device you plan to use. Some are desktop only.

How to test on a free trial

Run the same screen in two tools on the same day. Make it specific: US-listed stocks above a market cap floor, P/E below a level you choose, positive free cash flow. Write the criteria down first so both runs match exactly.

The lists will differ. Look at the tickers that appear in one result and are missing from the other, then find out why. Usually the cause is a definition or an update lag, and the glossary should explain it.

Next, pick three names and check their figures against the latest 10-K or 10-Q on EDGAR. Work out the P/E ratio yourself from reported earnings and the current price. A gap that the glossary explains, such as adjusted earnings, is acceptable. A gap nobody explains is a problem.

Then rerun the screen a day later and note which figures moved. That tells you the real refresh schedule. If you want a standard screen to reuse on every trial, the steps in how to use a stock screener give you one.

Red flags that end the trial

Walk away from a screener that will not say where its data comes from or when it was last updated. The same goes for one with no glossary, because a filter you cannot define is one you cannot rely on. Other signs: results that plainly break your own criteria, figures that disagree with filings with no stated reason, and a trial that hides the exact features you would be paying for. Check how to cancel before the trial ends. Some plans renew on their own.

Scoring what you tested

The weight-3 rows below decide whether a screener can answer your questions at all, so score them first and score them from what you saw on the trial, since the feature page claims full marks on everything. A low score on any weight-3 row is reason enough to pass, however good the total looks. Weight-1 rows are tiebreakers. If you are also weighing a bundle of reports and ratings, run the same exercise with the research subscription criteria and compare.