Complaint, workflow, competitor, demand, hidden: 5 mining lanes
The 143 tactic version of this is exhausting to hold in your head. This is the
compressed one: five lanes, nine sources each. What matters is not the list, it
is that the five lanes have wildly different signal quality and most people
treat them as interchangeable.
A. Complaint mining
G2, Capterra, community threads, X, App Store, Google Play, Trustpilot, YouTube
comments, Product Hunt.
Cheapest lane, and the most misleading. Writing a review costs nothing, so
the pile over-represents anger and says nothing at all about the quiet majority
who simply left. Use it for vocabulary rather than volume: the exact phrasing
people use becomes your landing page copy. Count a complaint only once it
repeats across two independent sources.
B. Workflow mining
Upwork, Fiverr, job boards, YouTube tutorials, written SOPs, Excel and Sheets,
Notion and Airtable, "manual process" searches.
The strongest lane, because every item cost somebody money or hours. A gig
posted a thousand times is a thousand purchase decisions. Its weakness is that
it is slow to read, and that it tells you what people currently do rather than
what they would put a card down for.
C. Competitor weakness mining
Pricing, feature requests, changelog, roadmap, support forum, integrations, API
limits, export limits, enterprise-only features.
The most precise lane. Everything in it is published by the vendor, so it is
verifiable rather than reported. The trap: a gap an incumbent left open for
three years is usually left open deliberately. Ask why before celebrating. Often
the answer is that the segment does not pay.
D. Demand mining
Search, trends, autocomplete, keyword volume, "alternative to", "best X for Y",
"X too expensive", "X does not support", "I wish there was".
The only lane that hands you a number before you build anything. Also the
easiest to fool yourself with, because volume without competition data is half a
measurement. One warning specifically about trends: it ranks velocity of
attention, which is news, and news decays inside two days.
E. Hidden SaaS mining
GitHub, browser extensions, Notion templates, Airtable bases, Google Sheets,
community bots, internal tools, open-source projects, freelance gigs.
Highest ceiling, highest failure rate. Every item is a product somebody
already built and chose not to commercialise. Sometimes that is because they
could not be bothered, which is your opening. Sometimes it is because the total
market is 300 people, which is not.
How they combine
One lane on its own is a hypothesis wearing a number. The sequence that has
actually worked here: D gives you the size, A gives you the words, C tells you
whether the space is defended, and B and E tell you whether anybody is already
paying to solve it by hand. Survive all five and it is worth a week. Survive
only A and you have found an angry review.
@bench_press_pr · 3d ago · 2 replies
The line about C that people skip is the instruction to ask why the gap is open.
I spent four months on an export feature a large product had refused to build. It turned out they had built it twice and killed it both times, because the support load was worse than the revenue. That was knowable from their own forum before I wrote a line.
Reply
Report
@nine_tabs_open · 3d ago
The counter-case exists though. Plenty of gaps stay open purely because the incumbent is large and slow, and that is most of what a small team can actually win on.
Reply
Report