RU EN

AJTBD Methodology

Product DNA — AI CPO Methodology

Product DNA (Discovery — Navigation — Acceleration) is AI CPO's proprietary methodology, synthesizing best practices from 10+ product frameworks:

  • Jobs Theory — understanding user tasks, job hierarchy, switch forces
  • Unit Economics — business math: LTV/CAC, cohorts, payback
  • RAT (Riskiest Assumption Test) — risk prioritization, quick hypothesis tests
  • Behavioral Economics — loss aversion, anchoring, System 1/2
  • Strategic Foresight — trend analysis, scenario planning
  • ABCDX Segmentation — segmentation by jobs, not demographics
  • GTM Strategy — channels, funnels, 5 awareness levels
  • Conversion Framework — landing page structure, AHA moment
  • Monetization Models — value-based pricing
  • Respondent Recruitment — 9 channels for finding respondents

Just as DNA contains the complete blueprint of an organism, Product DNA contains the complete map of your product: who the customer is, what job they perform, why current solutions fail, how much they're willing to pay, and the route from idea to working business.

AI CPO automatically builds your product's Product DNA through dialogue, external data analysis, and generation of 21 strategic artifacts.

Three Pillars of Product DNA

FrameworkWhat It StudiesUnit of Analysis
Jobs TheoryWhy people switch solutionsJob
Value-Based PricingHow to package and sell the transformationState transformation (Product DNA)
SaaS Unit EconomicsHow to measure whether the hypothesis worksMetric + signal
Unified Value Formula
Value = (Benefit − Investment × ~2) ÷ Expectations   [Jobs Theory]
Price = Value ÷ 10   [Value-based pricing (Nagle & Müller, 2018)]
Profit = (Price − CAC − Ops) × N customers   [SaaS unit economics]

Jobs-to-be-Done (JTBD)

JTBD views products as tools that people "hire" to accomplish specific tasks. Instead of "What features do we need?" JTBD asks: "What job is the user trying to get done?"

Classic example: people buy a drill not because they need a drill — they need a hole in the wall. But the hole isn't the end goal either. The real Job: "Hang a shelf to organize space and feel order at home." This includes functional (hang), emotional (feel order), and social (show guests) components.

Job Hierarchy

Each job exists at a specific level in the hierarchy. This isn't an abstraction — it's a concrete model for navigating the job graph:

LevelDefinitionExample (time tracking)Product Significance
Deep Need Fundamental need Financial independence Inspires Big Job — not solved directly
Big Job Major life/work outcome Build a stable freelance income Positioning at Big Job level = powerful strategy
Core Job Task for which the product is hired Accurately track time per project This is your core value. Start here
Small Job Sub-tasks within Core Job Start timer, categorize Product features solve Small Jobs
Micro Job Atomic UI interactions Tap the "Start" button in one tap UX optimization at Micro Job level

Job Types (Detailed)

TypeDefinitionIndicatorProduct Action
Functional Job Specific task to accomplish Clear action + measurable outcome Core product functionality
Emotional Job How a person wants to feel Verbs: feel, experience, be confident UX, communication tone, design
Social Job How a person wants to appear to others Verbs: appear, show, impress Sharing features, client reports, dashboards
Financial Job Financial outcome Money: save, earn, not lose ROI calculator, savings in numbers
Ordinary Job Direct action toward a result Clear action + outcome Solve better than competitors
Orientational Job Understanding, choosing, evaluating options Find / understand / choose / compare SEO, content marketing, guides, comparisons
Tax Job Arises from defects of the current solution Caused by a bug or friction, not the user's goal Kill through automation → powerful Push for switching
False Job No behavioral evidence Only stated intention, no time/money spent Filter out of roadmap — no real demand

Job Properties: Sequence, Frequency, Virality

Each job has three operational properties that determine product strategy:

PropertyWhat It DeterminesStrategic Significance
Sequence Position in the chain: what comes before and after, where the end-value is Early jobs = commodity (price competition).
Late jobs = value capture (quality competition).
Killable Jobs = jobs earlier in the chain that can be killed. Yandex.Rent killed inspection, deposit, contract — everything before "move in."
Frequency How often the job occurs Daily = habit, retention, subscription.
Monthly = needs reminders, engagement loops.
Once = one-time purchase, no retention.
Virality Is it performed with/for other people? Solo = paid traffic is the only path.
Shared = referrals, word-of-mouth.
Social = viral growth, K-factor > 1 possible.

Switch Formula (4 Forces)

People switch products (or start using a new one) when:

Canonical Formula
SWITCH happens when: PUSH + PULL > INERTIA + ANXIETY

PUSH = dissatisfaction with the current solution (intensity × frequency)
PULL = attractiveness of the new solution (desired outcomes)
INERTIA = cost/effort of switching (time, money, learning, habit)
ANXIETY = fear that the new solution won't work (perception of risk)

switchScore = Σ(push weights) + Σ(pull weights) − Σ(inertia weights) − Σ(anxiety weights)
switchScore > 0 → likely switches
switchScore ≤ 0 → likely stays
ForceDescriptionExampleHow to Amplify/Weaken
Push What repels from the current solution "Excel breaks, formulas get corrupted" Content: "Sound familiar?", amplify pain in marketing
Pull What attracts to the new solution "Saw a colleague's tracker — convenient" Demo, AHA moment, social proof, case studies
Inertia (Habit) Habit with the current solution "3 years on Excel — used to it" Data migration, training, smooth transition
Anxiety Fears about switching "What if I lose my data?" Guarantee, free trial, case studies, FAQ
Loss Aversion
Losses feel ~2x stronger than gains (Kahneman). This means: the new solution must deliver ≥ 2x value compared to the current one to overcome habit. If your product is only slightly better — the switch won't happen. The 1/10 rule (price = 1/10 of value) creates a 10x ROI that reliably covers loss aversion.

Important: the loss aversion coefficient is NOT fixed at 2.0x. Empirical range: 1.3x-2.5x depending on job category and segment.

Switch Formula Across Artifacts

19 Data Points

Each job in Product DNA is described through 19 data point types. AI CPO collects them from chat messages, files, interviews, and external sources. The more data points collected, the more accurate all artifacts become.

#Data PointJob LevelProduct DNA CategoryWhat's Extracted
DP01First thoughtCore/Big JobPush triggerFirst thought when the problem arose
DP02Triggering eventCore JobPushSpecific triggering event
DP03Current solutionCore JobInertiaWhat they use before switching
DP04Passive lookingBig JobEarly pushPassive search phase
DP05Active lookingCore JobPush escalationActive search phase
DP06First solutionCore JobPullFirst solution considered
DP07Decision criteriaCore JobPull/AnxietyCriteria used to decide
DP08Social proofCore JobAnxiety reductionWho they turned to for advice
DP09First useMicro JobOnboardingFirst usage experience
DP10Moment of valueCore JobPull confirmationAHA moment — when they realized the value
DP11What told othersBig JobLanguage/framingHow they described it to others (→ marketing language)
DP12Recurring contextSmall JobHabit formationContext of regular usage
DP13WorkaroundsSmall JobUnmet micro jobWorkarounds they invented
DP14AlternativesCore JobCompetitiveWhat alternatives they considered
DP15Desired emotionsDeep NeedEmotional jobWhat emotions they want to experience
DP16Budget / WTPCore JobEconomicBudget and willingness to pay
DP17Others involvedCore JobSocial jobWho else is involved in the decision
DP18Time horizonCore JobContextUrgency and time horizon
DP19Success definitionBig/Core JobOutcomeDefinition of success

Data Quality Tiers

TierDP CoverageInterview CountArtifact Reliability
GoldDP01-DP19 all≥ 5High — can build the product
SilverDP01-DP12≥ 3Medium — can test hypotheses
BronzeDP01-DP07≥ 1Basic — only for initial drafts
Experimental< 7 DPs1Low — all artifacts = hypotheses
Most Valuable Data Points
  • DP11 (What told others) — user language → positioning, offers, landing page
  • DP10 (Moment of value) — AHA moment → onboarding, demo, trial
  • DP16 (Budget/WTP) — pricing, unit economics
  • DP02 (Trigger) — when and how to reach users → GTM, advertising
  • DP13 (Workarounds) — unmet tasks → new features

Consideration Set

The Consideration Set is a model of which alternatives exist in the buyer's mind before the purchase. Key Product DNA insight: the Consideration Set forms before the trigger, not after. By the time of active search, the buyer already knows 2-3 options.

Consideration Set Structure
Current solution: what they use now, satisfaction level, switching cost
Known alternatives: what they know, perceived pros/cons, how they learned about them
Beliefs: "all CRMs are expensive", "open source = no support"
Fears: "migration will break everything", "the team won't adopt it"

Activating Knowledge — 3-5 phrases that update the Consideration Set and trigger a Switch WITH YOUR product. Extracted from DP11 and orientational search queries. This is the bridge between content marketing and adoption.

ABCDX Segmentation

ABCDX is a segmentation model based on the economic value of the segment, not demographics. Each segment is defined through Jobs, Execution Criteria, and context, then evaluated on 4 parameters.

B/C Boundary Decision Tree

The hardest part of ABCDX is the boundary between B (good) and C (marginal). Operationalized rule:

B vs C Decision Tree
Customer = B (good) if ALL conditions:
1. Core Job confidence ≥ 0.60 (silver tier minimum)
2. Switch score > 0 (will switch)
3. WTP ≥ product price
4. Decision autonomy = individual or influencer

Customer = C (marginal) if ANY condition:
1. Core Job confidence 0.30-0.59 (bronze)
2. Switch score ≤ 0, but Push > 0.5 (may switch later)
3. WTP < price, but > 50% of price
4. Committee decision required (long sales cycle)

The 80/20 Rule

80% of Resources on A+B
80% of revenue from A+B. 80% of support from C+D. This is a systemic trap: C/D shout louder because the product poorly solves their jobs — but solving for them requires the same resources as for A+B, with 5-10x lower return. The right strategy: filter C/D at intake.

RAT Framework (Detailed)

RAT (Riskiest Assumption Tests) is a framework for systematically validating hypotheses. Detailed description: RAT Tests.

RAT Process

  1. Identify — List all product assumptions (10-20)
  2. Prioritize — Score P×I for each (1-25)
  3. Design Test — Choose the fastest, cheapest validation method
  4. Execute — Run the test with clear success criteria (CONFIRMED / REFUTED / UNCLEAR)
  5. Decide — Pivot, persevere, or kill based on results

6 risk categories, P×I scoring, Quick Test library, and Risk Card format — see Build Phase.

How It All Connects: Jobs → Segments → Artifacts

The entire AI CPO methodology works as a unified pipeline:

Pipeline from Data to Decisions
1. Data collection: chat → fact extraction (5 dimensions) → 19 data points
2. Job formulation: facts → Job Statements → Job Graph (hierarchy + connections)
3. Segmentation: Jobs + Execution Criteria + context → segments → ABCDX
4. Switch analysis: Push/Pull/Inertia/Anxiety → switchScore per segment
5. Strategy: segments + switch → positioning (for A), offers (by awareness levels), price (3 dimensions)
6. Validation: assumptions → RAT (P×I) → Quick Tests → CONFIRMED/REFUTED
7. Build: Jobs + RAT → feature prioritization → MVP
8. Unit Economics: LTV/CAC/Payback by segment → go/no-go
9. GTM: awareness levels + channels + offers → first customers

Artifact Dependencies

ArtifactDepends OnUsed In
Pain MapFacts (pain)Job Statements, Segments
Job StatementsFacts (pain, audience)Job Graph, Positioning, VP
Job GraphJob StatementsSegments, Competitors, Features
Segments (ABCDX)Job Statements, FactsPersonas, Positioning, Pricing, Features
PositioningSegments, CompetitorsOffers, Landing, Pitch
RAT TestsAll facts (assumptions)Features, ABCDX grades
Unit EconomicsSegments, Pricing, ChannelsGrowth Strategy, Pitch

Validation Sequence (Three-Framework Synthesis)

Validation Stages
Stage 0: RAT — list risky assumptions, riskiest first [Product DNA]
Stage 1: 7-15 Discovery interviews — job graph, Switch story, budgets. End with a sales attempt [Product DNA]
Stage 2: Manual service — 5 paying customers for a manual version. State transformation for each [Manual service validation (Ries, 2011)]
Stage 3: Prototype/landing — test visitor → lead conversion. First funnel data [Lean Startup (Ries, 2011)]
Stage 4: A/B test — only after sufficient traffic. Test concept, not button color [Controlled experimentation (Kohavi et al., 2020)]
Key Principle
All three frameworks converge: don't trust opinions — trust behavior and money.
— Lean Startup (Ries, 2011): "First 5 paying customers manually → then write code"
— Jobs Theory: "Jobs are studied based on past spending — NEVER about the future"
— Behavioral vs stated preference (Kahneman, 2011): "User opinion ≠ user behavior"

Platform Evolution

AI CPO is a self-improving system. Three evolution mechanisms run automatically:

  1. Feedback Digest (daily, 3:00 AM) — aggregates feedback per artifact. AI identifies patterns: "users frequently complain about segmentation accuracy."
  2. Evolution Analysis (weekly, Mon 4:00 AM) — Claude Sonnet analyzes patterns and generates improvements for the backlog: new prompts, features, fixes.
  3. Chat Audit (weekly, Mon 5:00 AM) — Groq analyzes chat sessions: where users get stuck, confused, or abandon the workflow.

Results are available in the backlog (click "Backlog" in the chat). Most requested improvements are promoted to actual platform updates.

Custom Artifacts and Voting

If you need an artifact that's not in the standard set — request it. Other users can vote for requests. With 3+ votes, the request becomes a candidate for promotion to the standard artifact set.

Previous Credits & Pricing