This is not a test that produces a strategy once four words are typed in. The tool splits the analysis into two steps: record concrete factors and their evidence, then connect one internal factor with one external factor into an action. The final output is a four-quadrant canvas and a TOWS action table, not just four sticky notes.

S and W are the strengths and weaknesses inside the subject of the analysis; O and T are the opportunities and threats outside it. The same sentence can change category with a different subject, so write down the subject, the goal and the time frame first. The draft can be used for personal study, service design or small-organisation planning, not only for commercial competition.

1. 01Write the subject of the analysis and the period, for example “a two-month digital skills course pilot at one library”. Do not put ten years of institution-wide problems onto the same canvas.
2. 02Write one checkable fact or assumption per factor and record the evidence source. Mark the sample behind an interview finding, and do not write an unconfirmed partnership intention as an already secured resource.
3. 03Add cross strategies, choosing SO, WO, ST or WT, then choosing the matching internal and external factors. The tool checks whether the referenced types match the chosen cross.
4. 04Write each strategy as a next action with an owner, a date, a verification metric and a stopping condition. Deleting a referenced factor keeps the strategy but requires the link to be re-checked.
5. 05Check the four columns and the action table on the canvas; export PNG for communication, SVG to keep vector detail, Markdown to keep the whole draft, CSV for the currently selected table and JSON for a complete restore.

The formation of SWOT cannot reliably be reduced to a story of one person inventing it in one year. The historical research by Puyt and colleagues discusses its relationship to SRI's SOFT method and to strategic planning practice. This page therefore does not equate the involvement of Albert Humphrey or other figures with sole authorship. Historical research

Weihrich's 1982 TOWS paper proposed systematically matching the external environment with internal conditions. SO can use strengths to try to seize an opportunity; WO can use an opportunity to remedy a weakness; ST can use strengths to face a threat; WT can reduce simultaneous exposure to a weakness and a threat. These crosses are for proposing candidate strategies and do not guarantee that every combination is worth implementing. Original paper

Unlike a weighted decision matrix, SWOT does not apply one set of criteria to all options and computes no single score; unlike a risk matrix, it defines no event probability or impact level. To compare several actions, send a small number of clear candidates into a decision matrix rather than counting how many entries each column holds.

Hill and Westbrook's study of real consulting projects found long lists, vague descriptions, missing verification and results that never entered the subsequent strategy process in the cases they examined. That is evidence about particular practice, not proof that all SWOT is useless. This tool therefore emphasises sources and traceable actions, but it cannot carry out the investigation for the user. Study

The built-in “Community library course pilot” holds four factors and two strategies. S is volunteers who know the residents well, O is a community partner who may co-promote enrolment, and the SO action is to interview ten residents together before designing a trial class. The “ten” is an executable sample plan in the example and does not imply statistical representativeness. W is limited evening teaching capacity, T is other activities competing for the room, and the WT action is to book the room at weekends and cap the first intake.

A second example is a “change of major study plan”: it links insufficient statistical background with cross-department electives into a WO action — finish the preparatory unit, then sit in on a class. Another ST action uses a familiar internship domain to design a small project that tests the interest, rather than committing to the change of major straight away.

Factor S: volunteers know the residents → evidence: rostering interviews
Factor O: the community is willing to co-promote → evidence: meeting minutes, still to be confirmed
Cross SO: joint interviews → owner: course coordinator → date: 2026-10-10
Verification: record the specific needs; do not treat enrolment interest as paid demand

• The classification depends on the boundary of the analysis, and “quadrant completeness” cannot be compared directly across organisations.
• The canvas does not calculate budget, benefit, causal strength or probability, and it does not verify the content of external links.
• Up to 80 factors and 40 strategies; long text should be broken into explicit questions first.
• Complete text on the canvas can be long, PNG is limited by the browser's pixel budget, and JSON is preferred for a complete restore.

Treating “we hope for more funding” as an opportunity disguises a wish as an external fact. Classifying every problem as a threat hides the processes you can change yourself. “Improve quality” as a strategy does not say what will be done or who will verify it. It can be rewritten as “sample the last twenty delivery records, identify the types of repeated rework, and check again at the next delivery”.

Do not manufacture all four crosses to fill the table, and do not delete factors that argue against the current direction to get a tidy conclusion. Keeping the uncertainty and the stopping conditions is more useful than a symmetrical four columns.

• Suited to a clearly scoped course pilot, a club event, service improvement, a personal skills pivot and project kick-off discussions.
• When several clear options already exist, prefer a decision matrix; when a failure has already happened and its cause is being sought, prefer a fishbone or a root-cause chain.

Input, drawing and file generation happen in the browser. Local archiving can be cleared by the browser and is not a backup. A share link writes the document into the URL fragment, unencrypted, and anyone holding it can read the content. A link above 8 KB is not copied, and a JSON file should be used instead. Do not put personal identifying data, unpublished customer information or confidential strategy into it. A JSON import validates the tool type, fields, references and size, and a failure does not replace the current document.

• Weihrich, the original TOWS paper, 1982: the original methodology behind cross strategies. （访问日期：2026-09-22）
• Hill and Westbrook, 1997: empirical research and criticism of the quality of real-world use. （访问日期：2026-09-22）
• Puyt, Lie and Wilderom, 2023: a historical reconstruction of the origins, which does not treat the popular account as the only version of events. （访问日期：2026-09-22）