2025 H2 startup company ideas
Another half a year, more free ideas 2026-02-03 #startups

- 2014 ideas
- 2015-2017 ideas
- 2018 app ideas
- 2018 development ideas
- 2018-2019 ideas
- 2020 ideas
- 2021 ideas
- 2022 ideas
- 2023 business software ideas
- 2023 ideas (lifestyle)
- 2024 H1 ideas
- 2024 H2 ideas
- 2025 H1 ideas
- 2025 H2 ideas ←
I don’t attach any particular value to startup company ideas. But I continue to collect them, because I enjoy spotting problems that I wish someone (else) would solve, even if only for the entertainment value.
Food
Solve the one-true recipe problem: needing a new cookbook when you introduce a small variation in an existing recipe. For example, store all possible recipes and variants in a vector database of ingredients and cooking steps. Then you could discover and create similar recipes that only have small changes, and explore them by browsing higher-dimensional recipe space.
Solve the hangry cat problem: pet-feeding drama caused when putting the clocks back suddenly makes meals one hour ‘late’. For example, program automatic pet feeders to gradually phase out daylight savings time over the course of several weeks.
Reading
Solve the big book problem: unnecessarily-large non-fiction paperback books that have ridiculous amounts of padding, and blank pages. This especially affects business books with dozens of chapters, and 1–2 blank pages at each chapter transition. These books take too much space in your bag when travelling, and hurt when you fall asleep reading one and drop it on your face. For example, open a web shop that only sells A-format 110 × 178 mm mass market paperback reprints of books previously published in large trade paperback sizes.
Solve the book overflow problem, which occurs when young professionals live with their parents while saving to buy a house, and buy ‘too many’ books. For example, rent out 40-foot containers repurposed as cosy libraries: wall-to-wall book cases, wood panelling, armchair, reading light, rug; book storage to escape to.
Writing
Solve the pointless proof-of-work problem: people expecting a twenty-page document, when one page would do. For example, build a tool that takes a one-page ‘executive summary’ PDF, and appends twenty pages that appear corrupted, to make it look like you did the pointless proof-of-work.
Solve the cold text problem: colleagues who don’t know how to communicate effectively and emotionally online. For example, produce an app like Duolingo, but for learning how to use emoji (inline and reaction types) and memes (image macros and reaction GIFs) effectively. Sell downloadable content for specific industries and age groups.
Solve the contract origin problem: the taboo of source attribution in legal contracts, which lawyers treat as public domain works. For example, use stylometry and similar natural language analysis to trace the source of specific contract clauses .
Software
Solve the bug bankruptcy problem: software development teams’ huge backlogs unfixed bugs (related to the infinite backlog problem). For example, build a bug tracker that analyses each new bug report, based on past bug-fixing behaviour, and predicts the likelihood that you will ever fix the bug. It typically recommends that you don’t bother creating the bug report.
Solve the Excel date problem: Excel thinking everything’s a date. Make an otherwise-normal spreadsheet that only ever represents dates in an ISO 8601 date format, and never interprets anything else as a date.
Solve the SaaS museum problem: accidentally taking (and then hating) a job at a company that uses twenty year-old software tools, instead of modern alternatives. For example, create a job-board that lets you filter on usage of the most-loved and most-hated B2B software products.

