Skip to content
iSkllsSign up free

Shortlist a problem worth building

Group research notes into pains, pick the strongest one, and list how people search for it.

6 steps3 rulesStep 1 of 7 in Research a market

Use this skill

Read https://isklls.com/s/shortlist-a-problem and follow it.

Paste it into Claude, ChatGPT, Codex, Gemini or any agent that can open a link, then add your task.

Save.md
shortlist-a-problem.md
Raw

Purpose::

Find the problem worth building this week and the searches people type around it.

Rules::

  • Use only the notes you collected; the problem needs 2+ notes behind it
  • Skip anything you already built or chose
  • Search terms are 1-4 words, how people search for a product type; no sentences

Inventory::

  • Your research notes (posts, reviews, threads, cards)
  • Your list of past picks

Steps::

  1. 1Group pains and buyer words that mean the same thing
  2. 2Keep groups backed by 2+ notes and relevant now or within 2 years
  3. 3Pick the strongest pain; name it as a thing in a few words
  4. 4Write the problem: <who> struggle to <job> because <reason>
  5. 5List 10-15 terms: "<x> software", "<x> tool", "<x> template", "<x> for <who>", newer names
  6. 6Reply with Thing, Problem, Notes, the 2 runner-up groups and Terms

Metrics::

  • The thing traces back to 2+ notes
  • Every term could show search volume in a keyword planner
Research.md

Purpose:: Choose the search term to build for and show why it beat the rest.

Steps::

  1. 1. Get volume and cost per click for every term; score and sort
  2. 2. Top 3: read each 5-year trend as rising, steady, falling or spiky
  3. 3. Top 3: search each; note what ranks and where the gap is
  4. 4. Pick the winner, even if none is perfect

Pick the winning search term

Weigh volume, cost per click, 5-year trend and what already ranks, then pick one term and say why.

Research.md

Purpose:: Find the 5 strongest products that already ARE your idea, and gather what they do, so you build with proof.

Steps::

  1. 1. Write Is (the product), Is not (what looks close but isn't) and Looks like (its homepage hero)
  2. 2. Search your idea + up to 4 terms; collect unique domains
  3. 3. Open each homepage; quote its hero line; keep it only if it fits Is and not Is not
  4. 4. Search "<best direct> alternatives"; check each new name the same way

Find your 5 direct competitors

Find the 5 strongest products that already are exactly what you want to become, with proof.

Research.md

Purpose:: One table: each direct competitor's best feature, who it is best for, and its cons from real reviews.

Steps::

  1. 1. Per competitor: read the homepage; note its best feature and who it is for
  2. 2. Search "<domain> reviews" (the domain, not the brand name)
  3. 3. Read 1-2 review pages; note the cons people repeat
  4. 4. Reply with the table, one "Gap for us" line and one Sources line

Score competitors from real reviews

One table per market: each competitor's best feature, who it is for and its cons from real reviews.

Research.md

Purpose:: Given a URL, sign up and walk the real journey to first use, as reviewed screenshots of every step.

Steps::

  1. 1. Capture the landing page; find the main signup CTA
  2. 2. Sign up; handle cookie banners and popups yourself
  3. 3. Complete onboarding with reasonable demo choices; note every choice
  4. 4. Do one real free first-use task; capture the result and any upgrade prompt

Capture a website's signup funnel

Walk a site as a new visitor from landing to first use and save every step as reviewed screenshots.