Continuous Discovery Habits
A 10-14 week installation of Teresa Torres's Continuous Discovery Habits methodology layered with T2D3 proprietary IP for $5-30M ARR B2B SaaS teams stuck in the feature-factory pattern. Installs product trios, weekly customer interview cadence, Opportunity Solution Trees, assumption testing across the five risk categories, a research repository with tagging governance and traceability, and the Pain-Claim-Gain stakeholder operating mechanism. Reference methodology: https://www.producttalk.org/. Total scope: 7 modules, 28 sections, 62 tasks, 292 hours of canonical effort. Ongoing operating layer post-installation.
M1 - Trio Formation & Outcome Setting
Establish the smallest viable decision-making units (trios), pick a single product outcome per trio, and translate that outcome into an OKR. This module is the load-bearing first 2 weeks of the installation. OKR design captured per task: O1 (Become outcome-driven, KRs: interview cadence, OST coverage, assumption tests/opportunity, cycle time, repo coverage) and O2 (Improve product outcomes, KRs: activation lift, feature adoption, invalidation rate).
M1.S1 - Charter the Discovery Initiative
Run the executive pre-mortem on the feature-factory pattern and audit the last 5 features for outcome traceability. Establishes the leadership commitment that downstream cycle-time SLAs depend on.
Run pre-mortem on the feature-factory pattern with executive team
This pre-mortem is the leadership-alignment dependency for the entire Continuous Discovery installation. Teresa Torres' core diagnosis is that most organizations are…
Audit current discovery practice and feature-to-outcome traceability
This audit produces the baseline scorecard the entire playbook is measured against. Without it you cannot set realistic OKR targets, and you cannot demonstrate progress against…
M1.S2 - Form the Product Trios
Define the trio roster (PM + Designer + Engineer per outcome) using the T2D3 Syntropy capacity model and document trio rules of engagement (decision rights, conflict resolution, working cadence).
Define trio roster using the Syntropy capacity model
The product trio — PM + Designer + Engineer working together against one outcome — is the foundational unit of continuous discovery. The roster decision matters because Torres'…
Document trio rules of engagement and conflict-resolution playbook
Rules of engagement convert three individuals into one decision-making unit. The load-bearing principle from Torres is shared accountability across all five risk areas —…
M1.S3 - Define Product Outcomes
Translate the company-level business outcome into a trio-level product outcome and formalize O1/O2 plus their 8 KRs in the OKR system. Discovery cadence KRs become leading indicators.
Translate business outcome to product outcome (T2D3 worksheet)
A trio cannot directly move revenue or retention — those are business outcomes, lagging and influenced by forces outside product. Torres' fix is the product outcome: a change in…
Formalize O1/O2 + KRs in the OKR system
This task locks the playbook's two anchor objectives into the OKR system so discovery has teeth, not just intent. O1 (Become outcome-driven) measures whether the habits are…
M1.S4 - Stakeholder Alignment Kick-off
Build the Pain-Claim-Gain executive pitch deck and the stakeholder influence/impact matrix with per-stakeholder communication SLAs. Locks in HIPPO-intervention response protocol.
Build Pain-Claim-Gain executive pitch deck
The Pain-Claim-Gain deck is not a one-time pitch — it is the operating mechanism that defends discovery cadence under sprint pressure for the next 12 months. When a HiPPO demands…
Build stakeholder influence/impact map and per-stakeholder communication SLAs
Discovery dies when a high-influence stakeholder churns out of alignment and starts injecting feature demands mid-cycle. This map is the operating contract that prevents that. The…
M2 - Interview Recruiting Operations
T2D3 proprietary IP layer. The Torres methodology assumes interviewees show up; in practice, recruiting is the #1 reason continuous discovery fails. This module installs the recruiting machine: in-product prompts, CSM-referral SOP, panel-vendor protocol, scheduling automation, incentives policy, and GDPR/CCPA-compliant consent. Highest-leverage T2D3 IP layer in the playbook.
M2.S1 - Recruitment Channels
Stand up the four recruiting channels: in-product prompt (Pendo/Appcues/Sprig), CSM-referral SOP, panel-vendor contract (User Interviews / Respondent / Userlytics), and customer advisory community.
Build in-product recruit prompt (Pendo / Appcues / Sprig)
The in-product prompt is the bottom-of-funnel recruiting channel that keeps your weekly trio interview cadence alive when CSM referrals dry up and panel-vendor budgets get cut.…
Build CSM-referral SOP with weekly quotas and warm-intro template
CSM referrals are the highest-conversion recruiting channel in continuous discovery — 60–70% of warm intros convert to a scheduled interview, versus 10–15% for in-product prompts.…
Negotiate panel-vendor contract for top-up recruiting
Panel vendors are the third-tier recruiting layer, behind in-product prompts and CSM referrals. They exist for one job: top-up the funnel in the weeks your owned channels…
Stand up customer advisory community for repeated lightweight research
A customer advisory community of 30–50 opt-in members is the long-tail recruiting layer — the one that survives in-product prompt fatigue and panel-vendor cost spikes because…
M2.S2 - Incentives Policy
Define the B2C cash/credit/swag and B2B non-cash incentives ladder, and lock finance/procurement sign-off on the disbursement mechanism (Tremendous/Rybbon/direct payouts).
Define B2C incentives ladder and B2B non-cash menu
The incentives matrix exists to set one floor and one ceiling so individual interviewers don't over- or under-incentivize and skew the recruitment funnel. Without it, one PM…
Get finance/procurement sign-off on incentive disbursement mechanism
The single most common way a continuous-discovery recruiting pipeline dies in month 2 is procurement friction: if every incentive payout requires a one-off PO, the trio quietly…
M2.S3 - Scheduling Automation
Stand up Calendly + auto-Zoom-recording + transcript pipeline (Otter/Grain/Fathom) and define the no-show recovery flow with per-channel benchmarks.
Stand up Calendly + auto-Zoom-recording + transcript pipeline
Torres's non-negotiable for sustainable continuous discovery is automation: the goal is to wake up Monday with an interview already scheduled, doing nothing. Zero-touch interview…
Define no-show recovery flow with per-channel benchmarks
No-shows are the #1 way the weekly cadence dies in week 4. The trio books three interviews, one bails, and without a recovery flow the week's discovery collapses to a scramble —…
M2.S4 - Consent, Privacy & Quota
Build the GDPR/CCPA-compliant consent template, define PII retention + deletion policy, and lock in the cadence quota: >=1 customer interview per trio per week, exec-sponsored. Includes the EU/GDPR-gated consent flow.
Build GDPR/CCPA-compliant consent template
Continuous discovery generates a continuous stream of recorded customer conversations, transcripts, and PII flowing into a shared repository — which makes a documented consent…
Define PII retention + deletion policy
The consent template promises participants their data won't live forever — the retention + deletion policy is what makes that promise true. It's the mechanism that keeps the…
Lock in 1 interview per trio per week quota
In Teresa Torres' Continuous Discovery Habits, the single non-negotiable habit is weekly customer touchpoints — at least one interview per product trio (PM, designer, engineer)…
M3 - Story-Based Interviewing Skills
Train the trio in Torres's story-based interview craft. Most teams default to leading questions and feature-validation prompts; this module installs the muscle for 'Tell me about the last time you...' and the discipline to capture every interview as a one-page snapshot artifact.
M3.S1 - Interview Craft Training
Create the story-based interview guide template, run the 4-hour trio interviewing workshop with mock-interview replay, and define the engineer onboarding ladder (watch -> observe -> notes -> conduct).
Create story-based interview guide template
A story-based interview guide is the trio's go-to-meeting artifact. It exists so interviewers never improvise — improvisation is where leading questions, hypotheticals, and…
Run 4-hour trio interviewing workshop with mock-interview replay
This workshop is the only place the trio practices before the first real interview. Skip it and you set up the silo-relapse pattern: engineers default to feasibility-only…
Define engineer onboarding ladder for interview craft
Engineers are the trio members most likely to disengage from discovery — fear-of-the-unknown plus a strong feasibility instinct pulls them back to "just tell me what to build."…
M3.S2 - First Interviews & Snapshot Practice
Adopt the Product Talk one-page interview snapshot template and run the first 4 pilot interviews per trio with full attendance and snapshot completion within 24 hours.
Adopt Product Talk one-page interview snapshot template
The interview snapshot is the atomic artifact of continuous discovery. One interview produces one snapshot — a single page that captures who you talked to, what they said, and…
Run first 4 pilot interviews per trio with full attendance
The first 4 pilot interviews are the trio's proof-of-concept — the first time the PM, designer, and engineer conduct discovery together — and they generate the seed data for the…
M3.S3 - Bias Detection & Continuous Improvement
Build the interview replay rubric (closed-question count, leading-question flags, story-vs-summary ratio) and stand up the weekly trio interview retro. Surfaces craft drift fast at scale.
Build interview replay rubric with red-flag library
Interview craft decays silently. Without a scoring instrument, "that interview felt good" is the only feedback the trio gets, and quality drifts unnoticed until the OST fills with…
Stand up weekly 30-minute trio interview retro
Novelty carries interview quality for the first few weeks; then it fades, and without a forcing function the craft plateaus and drifts. The weekly 30-minute interview retro is the…
M4 - Opportunity Solution Tree (OST) Build
Install the OST as the trio's shared mental model and the bridge between discovery and roadmap. Covers tooling (Vistaly / Miro / FigJam), anti-pattern audits (solutions-disguised-as-opportunities, single-tree-per-trio), target-opportunity selection via importance x tractability, and the discipline of compare-and-contrast solution brainstorming.
M4.S1 - OST Tooling & Template
Select the OST tool (Vistaly / Miro / FigJam) and build the OST template with anti-pattern callouts and sample-tree examples drawn from Snagajob and FCSAmerica.
Select OST tool (Vistaly / Miro / FigJam) and configure trio workspace
The Opportunity Solution Tree only stays alive if it lives in a tool the trio touches every week. A tree drawn once in a slide deck dies in week two. The tool choice is a…
Build OST template with anti-pattern callouts and sample tree library
A blank canvas forces every trio to reinvent the Opportunity Solution Tree, and inconsistency compounds across trios. The template is the trio's shared mental model made into an…
M4.S2 - V1 Opportunity Space
Synthesize the first 4 interview snapshots into a v1 opportunity space (8-15 nodes) and run the anti-pattern audit to remove solutions-disguised-as-opportunities.
Synthesize first 4 snapshots into v1 opportunity space
This is the moment the trio's interview practice meets its first Opportunity Solution Tree. The job is to convert the 4 pilot interview snapshots into a v1 opportunity space — a…
Run v1 OST anti-pattern audit
The v1 Opportunity Solution Tree gets exactly one quality gate before it becomes the operating tree for the next six weeks — this audit. Skipping it is how trios end up with…
M4.S3 - Target Opportunity & Solution Brainstorm
Choose one target opportunity using the importance x tractability rubric (effort excluded) and brainstorm a minimum of 3 candidate solutions for compare-and-contrast.
Choose target opportunity using importance x tractability rubric
Picking the wrong target opportunity dooms the whole cycle: the trio spends six weeks testing solutions to a need that will never move the outcome. This task forces the trio to…
Brainstorm minimum 3 candidate solutions per target opportunity
Torres calls solution brainstorming the place "where most learning happens" — and the three-minimum rule is what makes it work. A single idea triggers a binary go/no-go and the…
M4.S4 - OST as Living Document
Establish the bi-weekly OST update working session and the monthly cross-trio OST review. Stops the tree drifting into a stale wishlist and surfaces duplicate opportunities across trios.
Establish bi-weekly OST update working session
An Opportunity Solution Tree is a living document or it is nothing. Left alone, it drifts into a stale wishlist within weeks — new interview snapshots pile up unsynthesized,…
Stand up monthly cross-trio OST review
Once 3+ trios run discovery in parallel, they silo. Each builds its own Opportunity Solution Tree, and without a forcing function they re-discover the same opportunities in…
M5 - Assumption Mapping & Testing
Surface and test assumptions across all 5 risk categories (desirability, viability, feasibility, usability, ethical) before building. Install the 12 lightweight test patterns and the cycle-time discipline that separates real discovery from theater. Achieves KR1.3 (>=3 tests per target opportunity) and KR1.4 (cycle time <=14 days).
M5.S1 - Assumption Surfacing
Build the 5-category assumption-test design rubric, adopt David Bland's importance x evidence grid, and run the 2-hour trio workshop to surface >=10 assumptions per candidate solution.
Build 5-category assumption-test design rubric
In Continuous Discovery Habits, once a trio commits to a candidate solution on the Opportunity Solution Tree, the next move is not to build — it is to surface and test the…
Adopt David Bland's importance x evidence grid
Surfacing assumptions is only half the job — the trio still has to decide which assumption to test first. Left to instinct, trios test the easiest assumptions (the ones they…
Run trio workshop to surface 10+ assumptions per candidate solution
The assumption-surfacing workshop is the load-bearing exercise that determines what an entire discovery cycle's testing program will look like. In Continuous Discovery Habits,…
M5.S2 - 12 Lightweight Test Patterns
Build the 12-pattern test library (prototype, survey, data mining, research spike, concierge, Wizard of Oz, fake door, smoke, painted door, 5-second, card sort, A/B) and the assumption-type -> recommended pattern decision tree.
Build 12 lightweight test pattern library
In Continuous Discovery Habits, the point of assumption testing is speed — small, cheap tests run weekly, not multi-month research projects. The blocker is setup overhead: if…
Build assumption-type to recommended test pattern decision tree
A 12-pattern library is only useful if the trio can pick the right pattern fast. Without a matching rule, trios default to the pattern they know best — usually a prototype or a…
M5.S3 - Run Assumption Tests
Run the first 3 assumption tests per trio against the riskiest assumptions (KR1.3 first cycle) and document the readout per solution with go/iterate/pivot/new-opportunity decision.
Run first 3 assumption tests against riskiest assumptions
This is the moment a trio stops talking about discovery and starts doing discovery. In Continuous Discovery Habits, the first three assumption tests — run against the three…
Document test results readout per solution
A test result that lives only in the trio's heads ends the discovery loop prematurely — stakeholders never see the invalidation evidence, and the org's faith in continuous…
M5.S4 - Cycle-Time Discipline
Operating cadence section. Build the cycle-time dashboard (max days between any two discovery activities) and run the first quarterly discovery retro. Two tasks - exempt from the 8-task floor per Decision 2.
Build cycle-time dashboard with max-day SLA
Counting interviews is easy to game — a trio can batch three interviews on a Friday and claim a "weekly cadence" while going a month between any other discovery activity. Torres'…
Run first quarterly discovery retro
A weekly cadence can run on autopilot and still go stale — trios hit a quality plateau in year two and never notice, because nothing forces them to step back and inspect how well…
M6 - Research Repository & Synthesis
T2D3 proprietary IP layer. Snippets, taxonomy, governance, the 'research repository trap' defenses, and traceability from interview -> snapshot -> opportunity -> assumption test -> shipped feature. Second highest-leverage T2D3 IP layer, drawing on Anderson's repository-trap critique and Dovetail's tagging governance.
M6.S1 - Repository Tooling & Schema
Select repository tool (Dovetail / Notion / EnjoyHQ) by first running stakeholder-needs interviews to avoid Anderson's research repository trap, and design the cross-linked entity schema.
Select repository tool with stakeholder-needs interviews
The repository is where continuous discovery either compounds or dies. Most teams reverse the order — they buy Dovetail or EnjoyHQ first, design the schema for the researcher,…
Design cross-linked repository schema
A repository becomes a knowledge system — not a notebook — the moment its entities are cross-linked rather than filed in folders. The traceability spine of the whole playbook…
M6.S2 - Snippet Tagging Taxonomy
Define the global tag set (pain points / motivations / behaviors / quotes / decisions), the project-tag set per outcome, and the taxonomy dictionary with change-control governance.
Define global tag set
In Dovetail's taxonomy framework, global tags are the small, centrally-governed vocabulary used across every project — the canonical layer. They are deliberately few and…
Define project tag set per outcome
The project tag layer is where each trio gets the granularity it needs without polluting the global five. Project tags extend the global vocabulary with outcome-specific patterns…
Build taxonomy dictionary with change-control governance
A tag set without a dictionary is just a list — and lists rot. Dovetail's taxonomy guidance (Bec Sareff-Hibbert's "talking tagging taxonomies", and the "four pitfalls" piece) is…
M6.S3 - Synthesis Operating Cadence
Operating cadence section. Build the snapshot -> repository pipeline with auto-tagging assist and run the monthly cross-trio affinity-mapping session in MURAL/Miro. Two tasks - exempt from the 8-task floor per Decision 2.
Build interview-snapshot to repository pipeline with auto-tagging
Repository coverage (KR1.5: ≥90%) lives or dies on the cost of getting a snippet in. When tagging is fully manual, the burden is too high — trios skip it under cadence pressure…
Run monthly cross-trio affinity-mapping session
When 3+ trios run discovery in parallel, each builds a private view of the customer — and they re-discover the same patterns in isolation for months before anyone notices. The…
M6.S4 - Insight Traceability
Build the insight-to-decision traceability log (each shipped feature linked to >=1 OST opportunity AND >=1 assumption test) and the OST-coverage report that measures KR1.2.
Build insight-to-decision traceability log
KR1.2 — ≥80% of shipped features tied to an OST opportunity — is only credible if it is measured continuously, not self-reported quarterly. Self-report inflates: a team that feels…
Build OST-coverage report (KR1.2)
The traceability log (previous task) is the raw daily surface; the OST-coverage report is its stakeholder-facing rollup. It answers one question on a weekly cadence: what fraction…
M6.S5 - Repository Socialization & Anti-Trap
Repository onboarding roadshow, exec-sponsor demo, per-team office hours, embedded Slack #research-snippets channel, and weekly highlight digest. Publishing/hand-off section - exempt from the 8-task floor per Decision 2.
Interview stakeholders on repository needs
The research repository trap, named in Anderson's critique, has a single root cause: the repository is designed for the researcher who built it, not the people meant to use it. "I…
Run repository onboarding roadshow
A repository's adoption curve is not automatic — it has to be driven. The needs interviews told you what consumers want; the onboarding roadshow is how you put it in front of them…
M7 - Stakeholder Co-Creation Cadence
Move discovery from project to BAU. Pain-Claim-Gain monthly stakeholder updates, Now/Next/Later roadmap tied to OST opportunities, OKR re-anchoring at quarterly planning, trio compensation/recognition (T2D3 IP gap fill that Torres explicitly leaves open), and the STOP-framework hand-off to BAU runbook covering all 22 playbook artifacts.
M7.S1 - Pain-Claim-Gain Stakeholder Updates
Operating cadence section. Build the monthly stakeholder-update template using the T2D3 Pain-Claim-Gain narrative walking the OST top-down, and deliver the first monthly update to executive team and board. Two tasks - exempt from the 8-task floor per Decision 2.
Build Pain-Claim-Gain monthly stakeholder update template
The monthly stakeholder update is the single operating mechanism that defends discovery cadence under sprint pressure for the next 12+ months. Most discovery initiatives die not…
Deliver first monthly Pain-Claim-Gain stakeholder update
The first update sets the tone for the next 12 months. Landing it well is the difference between genuine leadership buy-in and the quiet skepticism that erodes the cadence by…
M7.S2 - Outcome-Based Roadmap Integration
Build the Now/Next/Later roadmap with each lane traced to OST opportunities + assumption tests, and re-anchor product OKRs to validated opportunities at quarterly planning.
Build Now/Next/Later roadmap tied to OST opportunities
A Now/Next/Later roadmap is the artifact that converts the Opportunity Solution Tree into a stakeholder-readable narrative. Without it, the OST stays trapped inside the product…
Re-anchor product OKRs at quarterly planning
OKRs decay. Left alone, Key Results drift back into output metrics ("ship the redesign," "launch 3 features") within two quarters, and the whole outcome discipline quietly…
M7.S3 - Trio Compensation & Recognition
Sign-off section (T2D3 IP gap fill). Define the trio comp/recognition philosophy (reward outcome lift not feature ship-count) and the spot-bonus rubric for assumption-test wins. Two tasks - exempt from the 8-task floor per Decision 2.
Define trio compensation/recognition philosophy (T2D3 IP)
This is a T2D3 IP gap that Torres explicitly leaves open (producttalk.org/2020/06/measure-discovery): the discovery literature tells you to measure outcomes, but it stays silent…
Build discovery spot-bonus rubric
A philosophy that "rewards outcome lift" stays a slogan until it lands on an actual paycheck. The discovery spot-bonus rubric is the operating mechanism that operationalizes the…
M7.S4 - STOP Framework Hand-off to BAU
Publishing / hand-off section. Run the STOP hand-off checklist (Standardize -> Templatize -> Optimize -> Productize) across all 22 playbook artifacts and publish the trio operating-cadence BAU runbook. Two tasks - exempt from the 8-task floor per Decision 2.
Run STOP hand-off checklist across 22 playbook artifacts
The pilot is over; the question is whether the discovery cadence survives without it. The STOP framework (T2D3 IP: Standardize → Templatize → Optimize → Productize) is the…
Publish trio operating-cadence BAU runbook
The BAU runbook is the artifact that hands the discovery cadence to the organization as standing operating procedure — so it survives leadership churn. Without a documented…