The Two Paths to Opinionated Software
Either encode a best practice that already exists without technology, or teach a better way and enforce it in the product
- Difficulty
- Moderate
- Time to result
- ~months to results
- Steps
- 5
- Confidence
- 90%
Opinionated software has best practices or rules built into the system so that users can't work around them — the opposite of infinitely-customizable platforms. Gago identifies exactly two viable origins for it. Path one: find a process that is already a best practice (or a growing one) but has no technology enabling it, and build the tool around it — Atlassian with agile, and the devops tool wave. Path two: identify what companies are doing badly, form a strong opinion on the better way, teach it, and have the software enforce it — Greenhouse with structured recruiting, Pando with employee progression. The rules must be anchored in values, or the constraints feel arbitrary.
Origin
Barbra Gago, drawing on Atlassian (agile workflows), the devops tooling wave, Greenhouse's structured recruiting (which she brought to market), and her current company Pando's opinionated approach to killing performance reviews.
Core principles
- 01Opinionated software is principled software — the rules must be anchored in values, otherwise they are just friction.
- 02If the rules are anchored in values, certain behaviors become genuinely unworkable-around by design, and that is the point.
- 03Path one (encode an existing best practice) rides momentum that already exists; path two (teach a better way) requires you to fund the education.
- 04Enforcement earns its keep only where flexibility produces bad outcomes — bias, subjectivity, opacity.
How to run it
- 1
Decide whether the opinion should exist at all
Ask whether unconstrained flexibility in this workflow produces bad outcomes. Performance reviews are full of bias and subjectivity; unstructured recruiting is transactional and biased. Where flexibility is harmless, optimize for customization instead.
Watch out Enforcing rules on a workflow where users' custom approaches are fine just makes your product annoying.
- 2
Choose your path — ride a best practice or teach one
Path one: locate a process that is already a best practice or a growing one, has real momentum, and has no technology integrated — then build the enabling solution (Atlassian and agile; devops tooling). Path two: identify what companies are missing, define a better way, and commit to training and teaching them — then your software helps them do it (Greenhouse's structured recruiting).
Pro tip Path one is cheaper because the market has already validated the process. Path two carries category-creation-scale education costs.
- 3
Anchor the rules in values
Derive the specific enforced rules from a stated principle, not from product convenience. Greenhouse's structured recruiting — defined pipelines, interview kits — existed because it made recruiting better, more transparent, and less biased.
- 4
Build the enforcement into the system
Make the workflow require the behavior. Pando requires companies to be transparent about levels, expectations, and competencies, and creates accountability on both the employee and manager side — because without a system enforcing it, there is too much subjectivity for the process to be good.
Pro tip The honest pitch: you don't need our platform to do this — but if you don't do the work, you won't get the outcome. That framing makes enforcement feel like help rather than constraint.
Watch out Most companies won't do the work manually. That gap is exactly the wedge the opinionated product fills.
- 5
Pair enforcement with education
On path two especially, the software alone is insufficient — you must teach the better way through content and thought leadership, which is why opinionated software and category creation tend to travel together.
In the wild
Agile was already happening and had real momentum as a way of working, but had no technology integrated into it. Atlassian built the solution that enabled companies to do it with software.
→ An opinionated toolset for agile teams that succeeded because it encoded a practice the market had already validated, rather than having to teach it.
Rather than riding an existing practice, Greenhouse formed a view that there was a better way to recruit: a specific way to create a pipeline, build an interview kit, and structure the process. The rules were anchored in making recruiting better, more transparent, and less biased, and they were built into the product so they couldn't be worked around.
→ Structured recruiting became Greenhouse's differentiator, and — with Culture Amp working the adjacent space — helped set off a wave of stronger people practices across the industry.
Gago's view is that performance reviews perpetuate bias, and that companies patch the downstream symptom (pay gaps) by updating comp annually instead of fixing the systemic problem in how people are recruited, developed, evaluated, and paid. Pando therefore requires companies to be transparent about levels, expectations, and competencies, and enforces accountability on both employee and manager.
→ An opinionated system whose rules are derived from a value — employees should be evaluated fairly and paid fairly — rather than from product preference.
Common mistakes
Being opinionated without values behind the opinion
Rules that aren't anchored in a principle read as arbitrary constraints. Opinionated software works because it's principled — the value justifies the rule the user can't work around.
Taking the teach-a-better-way path without funding the education
Path two requires you to train and teach the market on a practice it isn't following. Shipping the enforcement without the content and thought leadership leaves users confused by constraints they don't yet believe in.
Is it for you?
Best for
B2B founders deciding whether their product should enforce a workflow or maximize configurability, especially in domains where subjectivity causes bad outcomes
Not ideal for
Horizontal platforms whose value is customization and flexibility, and domains where there is no defensible best practice to encode
From the transcript
“or maybe some rules integrated into the system you have systems that you can customize any which way and however you want and then you…”
“it's principled I would say it's anchored in some values and because of those values there will be certain rules that are built in that…”
“so I think there's two different approaches either find a process that just doesn't have technology that is a best practice or growing best practice”
“you don't have to use a platform to do it but if you don't do the work anyway you won't have the outcome that you…”
From the episode
Category creation and brand building
Barbra Gago (Pando, Miro, Greenhouse, Culture Amp)