How to Find SaaS Ideas: Rules, Job Posts and Review Complaints

To find SaaS ideas, look where money or obligation is already moving: a new regulation with a deadline, a job post for work software could do, or a review that names a missing feature. Each one is evidence that someone pays, or has to pay, to solve a problem. Log about 20 signals in a week, group them by buyer, and keep the ones that show up in at least two of the three sources. Then score what is left on pain, timing, revenue, market openness and ease before you write any code.

Why these three sources work

Brainstorming starts from what you can imagine. These sources start from what other people have already done: passed a rule, opened a requisition, or typed out a complaint. You are reading behavior, not guessing at it.

They also cover different questions. Regulation tells you why now. Job posts tell you what the work costs today. Reviews tell you what the current software gets wrong. An idea that answers all three is much stronger than one that answers only one.

Source 1: Regulation changes

A rule with an effective date creates a buyer with a deadline. Deadlines shorten sales cycles, because "we'll look at this next quarter" stops being an option.

How to do it

1. Pick a regulated industry you can reach: food service, trucking, property management, healthcare, childcare.
2. Check the Federal Register. Its Public Inspection page previews documents scheduled for the next day's issue, so you see rules before they are official.
3. Note the compliance date, who has to comply, and what they will have to do differently (record, label, report, test, file).
4. Ask: is the new work repetitive, and does it fall on small operators who have no compliance staff?

Two real examples

Sesame became the ninth major food allergen under the FASTER Act, which was signed in April 2021 and took effect January 1, 2023. Every kitchen that handles sesame had to rethink its labels. That is the kind of change behind an idea like allergen labels for ghost kitchens.

In trucking, FMCSA's Clearinghouse II rule had a compliance date of November 18, 2024, when state licensing agencies began downgrading CDLs for drivers with a "prohibited" status. One compliance guide notes that over 175,000 drivers held that status as of July 2024. Small fleets suddenly care about testing and scheduling, which is the territory of DOT drug test scheduling for small fleets.

A rule alone is not an idea. Large compliance vendors may already cover it, and a one-time change may not support a subscription. You want rules that create work every month.

Source 2: Job posts

A job post is a company saying, in writing, that it will pay a person to do a task. If the task is narrow and repetitive, that is a software candidate. The salary is a rough ceiling on what the problem is worth to that buyer.

How to do it

1. Search a job board by role or keyword. Indeed offers filters such as date posted, location and job type, which help you narrow to recent listings in one niche.
2. Skip the headline title and read the duties. Look for phrases like "track," "reconcile," "chase," "schedule," "enter," and "send reminders."
3. Count repeats. One post is an anecdote. Ten similar posts across different small employers is a pattern.
4. Check posting dates. Some listings are stale, and tools exist specifically to expose "ghost jobs", so do not treat every post as live demand.

Imagine ten home care agencies each posting for a coordinator whose duties are mostly phoning caregivers to cover open shifts. That would point toward something like shift swaps for home care. A paid role means the pain is expensive enough to staff.

Source 3: Review complaints

Reviews of existing software show you demand that has already been proven, plus the specific gaps. This is the cheapest source to mine because the customer has done the writing.

How to do it

1. Choose a category where small businesses buy software.
2. Open the top three products on a review site. G2 asks reviewers what they like best and what they dislike, so the complaints are already sorted for you. It also labels reviewers by company size, such as Small-Business or Mid-Market.
3. Read only the "dislike" answers from small-business reviewers. Tally the repeated complaints in a spreadsheet: price, missing integration, clunky setup, no support for a language or a trade.
4. Look for complaints that are about a segment the tool ignores, not about bugs. Bugs usually get fixed; a segment the vendor does not serve is a more durable gap.

A hypothetical pattern: if small-restaurant reviewers of a reply-management tool kept saying it only handles English, that would point toward something like Spanish review replies for local restaurants. The complaint names the segment and the gap in one sentence.

Be honest about the limits here. Review sites skew toward products that actively ask for reviews, and the loudest complaints are not always the ones that would make someone switch.

Combine the signals

Two sources agreeing is a decent filter. Three is rare and worth a closer look. Always list competitors, even if you think the space is empty. If nobody sells to this buyer, ask why before you assume it is an opening.

Score the shortlist

Proofline weighs five factors: pain at 0.30, why-now at 0.25, revenue at 0.20, open market at 0.15 and ease at 0.10. The total is then multiplied by an evidence factor between 0.70 and 1.00, so a well-argued idea with thin proof gets marked down. More detail is on the methodology page.

Here is a hypothetical run, with made-up 1-to-10 ratings used only to show the arithmetic. Pain 8, why-now 9, revenue 6, open market 5, ease 6. That works out to 2.4 + 2.25 + 1.2 + 0.75 + 0.6 = 7.2. With an evidence factor of 0.85, the final figure is about 6.1. Notice that a rule change can lift why-now, but it does not rescue weak revenue.

If you would rather see finished examples than build the spreadsheet yourself, today's scored idea and the idea archive show the same method applied with comparables. If you are weighing it against other idea databases, the comparison pages go through the differences.

Common mistakes

• Stopping at the signal. "A new rule exists" is not an idea. Name the buyer, the task and the price they would pay.
• Collecting only from one source. Single-source ideas are the ones that fall apart in the first customer call.
• Ignoring incumbents. A crowded review page means demand. It also means you need a reason to win, such as a niche, a price or a workflow.
• Picking industries you cannot reach. If you cannot get ten buyers on the phone, the best signal in the world will not help.
• Counting posts, not patterns. Ten posts from one large employer is one data point.
• Skipping ease. A good problem you cannot ship in a few months may lose to a smaller one you can.

A weekly routine

• Monday: scan the Federal Register and your industry's regulator for new compliance dates.
• Tuesday and Wednesday: read 30 job posts in one niche and log repeated tasks.
• Thursday: read "dislike" reviews for the three leading tools in that niche.
• Friday: merge, score, and talk to two people who hold the job title you saw.

For niche-specific lists to start from, see trucking, restaurants and real estate.