The gap between a SaaS product that struggles to retain users and one that people rave about is almost always research. The teams building the best SaaS products aren’t guessing — they’re systematically learning.
But not all research methods are equal. Some are slow and expensive. Some produce data that looks rigorous but doesn’t actually inform decisions. This guide cuts through the noise to show you the 10 research methods that consistently produce actionable insights for SaaS product teams.
The Research Methods Spectrum
Before diving in, it helps to understand two dimensions of research:
- Qualitative vs. Quantitative — Qualitative tells you why; quantitative tells you what and how much
- Generative vs. Evaluative — Generative research helps you discover new ideas; evaluative research tests ideas you already have
The best research programs use all four quadrants. Here are the 10 methods you need:
The 10 Essential SaaS Research Methods
1. User Interviews
Type: Qualitative, Generative/Evaluative | Time: 30-60 min per session | Effort: Medium
One-on-one conversations with users or potential users. The gold standard for understanding motivations, mental models, and pain points. Interview questions should probe the problem space (“Tell me about the last time you…”), not validate your solution (“Would you use this feature?”).
When to use: Before starting design, when you’re confused about why users behave a certain way, when usage data shows something unexpected.
2. Usability Testing
Type: Qualitative, Evaluative | Time: 30-60 min per session | Effort: Medium
Observing users attempting to complete tasks in your product. Reveals where users get stuck, what they misunderstand, and what they expect to happen. The “think aloud” protocol — asking users to narrate their thought process — is particularly valuable.
When to use: Before shipping a new feature, when you see drop-off in a specific flow, when you’re evaluating design alternatives.
3. Heuristic Evaluation
Type: Qualitative, Evaluative | Time: 2-4 hours | Effort: Low
Expert evaluation of your UI against established usability principles (heuristics). No users required. Trained evaluators can surface 60-75% of usability problems in a fraction of the time of user testing.
When to use: Early in the design process, before usability testing (to clear obvious issues), when you need fast feedback with limited resources.
4. Analytics Review
Type: Quantitative, Evaluative | Time: Ongoing | Effort: Low (once set up)
Analyzing product usage data to understand behavior patterns. Which features are used most? Where do users drop off? What does the path to activation look like? Tools like Mixpanel, Amplitude, and Heap make this increasingly accessible.
When to use: Always. Analytics should be a continuous background process, not a one-off activity.
5. Competitive Design Analysis
Type: Qualitative, Generative | Time: 4-8 hours | Effort: Medium
Systematically studying how competitors and adjacent products have designed their UI. Reveals category conventions (what users expect), design patterns worth borrowing, and gaps in the market you can exploit.
When to use: Before designing a new feature area, when entering a new market, when you’re designing your first version of something.
6. Survey Research
Type: Quantitative, Generative/Evaluative | Time: 1-2 weeks | Effort: Low
Collecting structured responses from many users at once. Great for measuring sentiment, prioritizing features, and understanding user demographics. The quality of survey data depends heavily on question design — avoid leading questions and keep surveys short.
7. Session Recording Review
Type: Qualitative, Evaluative | Time: 1-2 hours | Effort: Low
Watching real user sessions recorded by tools like Hotjar, FullStory, or LogRocket. You can see exactly where users click, where they hesitate, and where they rage-click in frustration. Extremely efficient way to discover usability problems without scheduling user sessions.
8. Card Sorting
Type: Quantitative/Qualitative, Generative | Time: 15-20 min per participant | Effort: Medium
Participants organize content into groups that make sense to them. Used to design or evaluate information architecture and navigation. Open card sorting generates new groupings; closed card sorting tests existing ones.
9. A/B Testing
Type: Quantitative, Evaluative | Time: 1-4 weeks | Effort: High
Running two versions of a design simultaneously and measuring which performs better against a specific metric. Requires significant traffic to produce statistically significant results. Best used when you have a specific hypothesis to test, not as a discovery tool.
10. Jobs to Be Done Interviews
Type: Qualitative, Generative | Time: 45-90 min per session | Effort: Medium
A specialized form of interview focused on understanding the “job” a user is hiring your product to do. JTBD interviews uncover the functional, social, and emotional dimensions of what users want to accomplish — and the forces that caused them to seek a solution in the first place.
Build a Reference Library for Your Research
Research is most powerful when it’s cumulative. The teams that consistently build the best products don’t start from scratch with each project — they maintain a library of design research that informs every decision.
The SaaS Boat library gives your team access to 3,600+ screenshots, animations, heuristic evaluations, and diagrams from the world’s best SaaS products — so your competitive analysis starts with real data, not a Google Images search.
