Pricing research from public content: what people already spend

The short answer: public content will not hand you a price, but it reliably gives you the two inputs that decide one — what your buyers already pay for adjacent tools, and which alternative they mentally compare you against. Get the comparison set wrong and no amount of price testing rescues you, because you are optimising inside the wrong bracket. As of August 2026 that comparison set is the most under-researched input in most early pricing decisions.

The usual approach is to look at three competitors, pick a number slightly below the middle one, and adjust later. It fails quietly, because the competitors a founder watches are frequently not the ones buyers weigh them against.

Reference prices are volunteered constantly

People discussing tools in public name numbers without being asked. "We were paying about eighty a month for that and dropped it." "It is twenty per seat, which for a team of six adds up." "Honestly at nine dollars I stopped thinking about it."

Each of those is a data point of a type surveys cannot produce, because the speaker had no reason to posture. The third is the most useful shape: it identifies the threshold below which the purchase stops requiring a decision, which for self-serve products is often more important than the maximum anyone would pay.

Note the fourth row is the one founders most enjoy finding and it is the weakest signal on the list. It costs nothing to say and predicts almost nothing, which is the same reason survey-based validation underperforms — the argument in validating a SaaS idea without surveys.

Finding the comparison set that actually applies

Price is judged relatively, and the reference object is chosen by the buyer, not by you. This is the highest-leverage thing to research, because it operates one level above the number.

The same product can be compared against a $9 utility, a $60 category tool, or half a day of a contractor's time. Those three framings support wildly different prices for identical software. What decides which one applies is how the buyer describes the alternative when they talk about the problem, and they do describe it — usually in the same breath as the complaint.

The last item deserves respect. Doing nothing is a live competitor with a price of zero and a strong incumbent advantage, and it wins more deals than any named rival. If your research shows people mostly absorb the problem, your pricing argument has to be against inertia, which is a different pitch from being cheaper than a competitor.

Reading price complaints correctly

A competitor being widely called expensive is tempting to read as an opening to undercut. It is more often a signal about legibility than about the number.

The check is whether the same buyers pay comparable or higher amounts elsewhere without complaint. If they do — and they usually do — then the problem is not the price level, it is that the value is diffuse and hard to point at. Entering at a lower price against an illegible value proposition inherits the same problem with less revenue to solve it.

The productive response is to find what the complainers do consider worth paying for, which is visible in what they defend rather than what they attack. A tool people describe as expensive but keep is priced correctly and marketed poorly; a tool people cancel is priced above its perceived value. Those are opposite situations that produce the same complaint vocabulary, and the teardown approach in running a competitor teardown from public content is the fastest way to tell them apart.

Pricing against a free alternative

In most categories a free option exists and shapes expectations whether or not it is genuinely comparable. Ignoring it does not work; arguing against it in the abstract does not either.

What works is naming the cost of free in units the buyer already uses. Not "our tool is more powerful", but the specific hours, specific risk, or specific rework the free path involves — the numbers your research already produced when people described their workarounds. That converts an apples-to-apples price comparison into a cost-of-the-whole-approach comparison, which is the only version you can win.

This is the same structure as pricing any tool against manual effort: the subscription is small and the labour is large, and the argument only lands when both are quantified honestly, including the cases where the manual path is genuinely the better call. Worked through in what YouTube research actually costs.

A workable research pass

The pass is narrow and produces a bracket rather than a number.

• Assemble sources where buyers discuss the workflow, not where vendors discuss the category. Reviews, walkthroughs, and workflow content beat vendor comparison videos.
• Extract every stated number with its context — what was being paid for, at what team size, and whether it was kept or cancelled. A price without those three is not usable.
• Record every alternative mentioned, including manual processes and doing nothing. This becomes the comparison set.
• Separate kept from cancelled. The gap between what people keep and what they drop is your bracket.
• Note the unit people use to describe volume — seats, clients, projects, months.

Fifteen to twenty sources is normally enough for the numbers to stop moving, and the same saturation logic applies as in any corpus, laid out in how many videos a research corpus actually needs.

From bracket to price

Research narrows; testing decides. What the pass gives you is the range your buyers already operate in, the comparison object you will be judged against, and the unit to charge in. Inside that bracket the remaining question — where exactly to land — is answered by putting a number in front of people and watching what happens.

Its value is in preventing the category-level error: pricing a $60 problem at $9 because you compared against utilities, or the reverse. As a live example, the tiers on this product sit at $19, $59 and $199 per month with 2, 8 and 20 concurrent projects, because the binding constraint buyers describe is how many research lines they keep open at once, not how many videos they watch — a unit that came out of exactly this kind of listening. The tier breakdown is here, and repeating the pass as the category shifts is the standing-research habit described in monitoring a niche instead of researching it once.

Closing thought

Pricing feels like it should be a modelling exercise, which is why it so often gets done in a spreadsheet with no external inputs. The inputs exist and they are free — people announce what they pay, what they dropped, and what they compare things to, constantly and without being asked. The work is not extracting a number from them. It is finding out which bracket you were in all along.