Drip
Case StudiesProcessCareers
Conversion Optimization LicenseCRO Audit
BlogResourcesArtifactsStatistical ToolsBenchmarksResearch
Book Your Free Strategy CallBook a Call
DRIP.
/PRIORITIZATION
DOSSIER 03 · SUCCESS RATE
BEHAVIORAL SCORING ACTIVE
THE THIRD VARIABLESUCCESS RATE · COMPOUNDING

The hardest question
in CRO isn’t how to test.
It’s what to test next.

Most teams answer it with opinion, or with a simple impact-and-effort guess. We score every hypothesis across several behavioral models first — so the test at the top of the list is the one most aligned with how your buyers actually decide.

▷ THE GROWTH PROTOCOL · FORMULA
ΔRPU=Quality×Rate×Success rate
THIS PAGE OWNS THE THIRD MULTIPLIER · §07 SHOWS THE FULL SYSTEM
▷ RANKED BACKLOGTOP 3 SHIPS THIS SPRINT
FOGG × BIG5 × DRIVERS
01
Digestive-outcome proof on PDP
Security · Progress
0
02
Founder video in hero
Trust · Belonging
0
03
Real-results carousel above fold
Security
0
04
Bundle suggestions in cart
Progress
0
05
Sticky add-to-cart on mobile
Convenience
0
06
Money-back guarantee banner
Security
0
07
Reorder benefits hierarchy
Clarity
0
08
Subscription default selected
Anchoring
0
09
Free shipping bar at top
Anchoring
0
10
Comparison table vs competitor
Status
0
11
Press logos band under hero
Authority
0
12
Reduce checkout to 3 fields
Convenience
0
— BELOW: WHY BEHAVIORAL SCORING BEATS BUSINESS SCORING
SECTION 01 / 08
● 02 · THESIS · WHY BEHAVIORAL SCORING WINSTHREE PARAGRAPHS · ONE MOTION

Most teams score the wrong thing.
We score what actually moves buyers.

Impact-and-effort tells you what a test is worth and what it costs. It tells you almost nothing about whether it will actually win.

I
Three variables drive a CRO program.
How good the test ideas are, how fast you can run them, and how often you pick the right one to run next. The first two get all the attention. The third — success rate — is the one almost nobody systematizes. And it’s the one that compounds.
II
Business scoring is better than gut. Still wrong.
Most teams that try use impact, confidence, effort. It beats arguing in a meeting. But impact and effort tell you what a test is worth and what it costs. Whether it wins is a question about human behavior — not about business math.
III
So we score on behavior instead.
Every hypothesis runs through several psychological models before it gets a priority. Each model produces a sub-score; the sub-scores combine into one number. The test at the top of the list isn’t the one with the biggest theoretical upside — it’s the one most aligned with how your buyers actually decide.
▷ THE WHOLE PAGE IN ONE MOTIONMANY MODELS · ONE SCORE · LOOPING
FoggB = MAP
Big FiveOCEAN
Drivers7 motives
Expert revieweffort + tags
Quantum priors4,127 tests
Iterative loopself-tuning
HYPOTHESIS · #H-01
Digestive-outcome proof on PDP
M·85A·60P·70Fit·78D·87
84
PRIORITY · COMPOSITE
SECTION 04 ZOOMS IN ON HOW EACH MODEL SCORES A REAL HYPOTHESIS
— NEXT: HOW MOST TEAMS DECIDE WHAT TO TEST
SECTION 02 / 08
● 03 · THE WRONG WAY · FOUR FAILURE MODESWHAT BEHAVIORAL SCORING REPLACES

Before a scoring system,
here’s how the queue gets built.

Each of these feels reasonable in the moment. Together, they’re how teams burn dev hours on low-impact tests, miss early wins that should have compounded, and eventually decide “CRO doesn’t really work for us.”

TOP OF Q
I

Loudest voice wins

The senior person, or whoever argues hardest, gets their test pushed to the front. Score by org chart, not behavior.

QUICK BUILD · 1HRSHIP NOW
II

Easy-first

Whatever's quickest to build jumps the queue. The roadmap becomes a list of one-hour wins that don't move anything.

N · 247UNSCORED
III

Backlog burial

Ideas pile into an unscored list, sit there, age out. The good ones get buried under the new ones.

COMPETITOREXIT-INTENT POP-UP10% OFFEXIT POP-UPYOUR ROADMAP
IV

Competitor copy

A pop-up on a competitor's site goes straight to the top, with no analysis of whether their buyers are anything like yours.

Even teams that upgrade to impact-and-effort are still guessing at the one thing that matters most:will this actually change behavior?
→ ANSWERED IN §04
— NEXT: HOW A HYPOTHESIS GETS SCORED · THE CENTERPIECE
SECTION 03 / 08
● 04 · THE CENTERPIECE · BEHAVIORAL SCORING CARDONE HYPOTHESIS · FOUR MODELS · ONE NUMBER

How one hypothesis
gets a real priority.

Watch the same idea pass through four behavioral lenses. Toggle the buyer profile to see sub-scores shift, the composite move, and the whole backlog rerank. That last part is what most teams miss: priority is a function of who’s buying.

◇ BUYER PROFILE · CHOOSE WHO YOU’RE OPTIMIZING FOR
Security-seeker. Wants proof before commitment. Reads everything. Hates ambiguity.
▷ HYPOTHESIS · H-01PDP · ABOVE FOLD
STATEMENT
If we place digestive-outcome proof above the fold on PDPs, then add-to-cart will increase, because this audience pre-commits only when the specific outcome is evidenced.
A · CONTROL
B · VARIANT
OUTCOME PROOF
COMPOSITE PRIORITY
FOR · SECURITY-SEEKER
0
OF 100
Fogg Behavior
B = M × A × P
Motivation84
Ability62
Prompt71
Big Five (OCEAN)
Trait alignment · top loads
Conscientiousness88
Neuroticism74
7 Core Drivers
Motive · which engine fires
Security92
Belonging58
Fit + Effort + Priors
Brand · build cost · history
Brand88
Effort74
Priors82
▷ TOP 5 · RERANKED BY PROFILE
The queue changes when the buyer changes.
SAME 20 HYPOTHESES · DIFFERENT WEIGHTS
#0184
Digestive-outcome proof on PDP
Security · Progress
#0278
Money-back guarantee banner
Security
#0374
Real-results carousel above fold
Security
#0471
Founder video in hero
Trust · Belonging
#0565
Press logos band under hero
Authority
▷ COMPOSITE FORMULApriority = w₁·Fogg + w₂·Big5 + w₃·Drivers + w₄·Fit · weights tuned per buyer profile
— NEXT: THE SCORE ISN'T STATIC. THE LOOP TUNES IT.
SECTION 04 / 08
● 05 · THE LOOP · WHY THE SCORE GETS BETTERPREDICT · SHIP · MEASURE · TUNE

The scoring system
tunes itself.

A static priority model is just a more confident guess. Every test result feeds back into the weights — over a quarter, the score stops being a hypothesis and starts being a calibrated prediction of what your specific audience responds to.

▷ THE LOOPCONTINUOUS
PREDICTSCORESHIPTEST LIVEMEASURERESULTTUNEREWEIGHTSELF-TUNINGEVERY SPRINT
Every win, loss, or flat test gets fed back. The weights on Fogg, Big Five, drivers, and brand-fit move toward whatever’s actually predictive for this audience — not toward a universal best-practice.
▷ MODEL CALIBRATION · 8 SPRINTSPREDICTED → ACTUAL · CONVERGING
255075100S1S2S3S4S5S6S7S8ACTUAL · WIN RATEPREDICTED · MODEL
S1 WIN RATE
38%
S8 WIN RATE
84%
GAP (PREDICT vs ACTUAL)
closes 24pt
"

Most CRO programs spend a year arguing about what to test. Ours spends that year learning what to weight.

DRIP — OPERATING PRINCIPLE 03
— NEXT: WHAT THIS LOOKED LIKE FOR ONE CLIENT
SECTION 05 / 08
● 06 · PROOF · ONE CASE, NINE MONTHSOCEANSAPART · DIRECT-TO-CONSUMER APPAREL

What happens when
the queue actually reranks.

The team had been running impact-and-effort. Many tests, all reasonable, win rate near coin-flip. After rewriting the priority model around behavioral scoring, RPU compounded for nine months straight.

▷ REVENUE PER USER · INDEXED 100 AT M01
Five tests, picked by the model. Four wins, one flat.
WINFLAT / LOSS
100120140BASEOutcome proof PDP+3%Founder video hero+4%Comparison block—Bundle in cart+6%Mobile sticky ATC+5%M01M02M03M04M05M06M07M08M09RPU · INDEX
0%
RPU lift
9 MONTHS · CUMULATIVE
0%
Test win rate
UP FROM 41% PRIOR PROGRAM
5 / 6
Sprints with a win
ONE FLAT, ZERO LOSSES
0.0×
ROI vs prior quarter
EVEN AT SLOWER TEST CADENCE
▷ WHAT CHANGED

The backlog stayed the same. The order didn’t.

01
Reprofiled buyers
Voice-of-customer + checkout exit surveys identified a security-driven core.
02
Re-weighted the model
Conscientiousness, Security driver, and outcome-evidence priors got upweighted.
03
Reranked the existing backlog
Three tests already in the queue moved from #14, #11, #9 → #1, #2, #3.
04
Shipped in that order
Each early win compounded into the baseline before the next test ran.
▷ FROM THE CLIENT
“The tests weren’t new. The order was. We’d have run the comparison block in month two and called the whole program flat. Instead we ran outcome-proof first, and everything after it inherited the lift.”
OA
Head of E-commerce
OCEANSAPART · APPAREL
CASE OPEN
Read the full case →
— NEXT: HOW THIS FITS THE WHOLE GROWTH FORMULA
SECTION 06 / 08
● 07 · THE WHOLE SYSTEM · WHERE THIS SITSTHREE VARIABLES · ONE FORMULA

Pick the right test next,
and the other two compound.

Prioritization isn’t a standalone product — it’s the multiplier in the growth formula. Stronger hypotheses without good prioritization just means a faster roadmap of mediocre tests. The three only work together.

▷ THE GROWTH FORMULA
ΔRPU=Quality×Rate×Success rate
YOU ARE HERE · SUCCESS RATE
VAR · 01
Test quality
STRONG HYPOTHESES, WELL-DESIGNED
4,127 prior tests indexed
Behavioral primitives, not surface ideas
Hypothesis template enforces clarity
§01 · QUANTUM DATABASEsee system →
VAR · 02
Test rate
SHIP MORE, LEARN FASTER
Up to 6 concurrent tests, multi-arm
Audience-isolated, no cross-contamination
Sequential analysis · early-stop logic
§02 · RAPID A/B TESTINGsee system →
VAR · 03YOU ARE HERE
Success rate
PICK THE RIGHT ONE TO RUN NEXT
Behavioral scoring · Fogg · Big5 · Drivers
Buyer-profile-weighted compositing
Self-tuning every sprint
§03 · ITERATIVE PRIORITIZATIONTHIS PAGE
▷ THE COMPOUNDING MATH · WHY ALL THREE MATTER
SCENARIO A · STRONG QUEUE, LOW SUCCESS
0.8×1.0×0.40=0.32×
A team running mediocre tests fast and well-managed. Output stays modest.
SCENARIO B · RIGHT IDEAS, WRONG ORDER
1.4×1.2×0.45=0.76×
Behaviorally rich hypotheses, but run in a coin-flip order. Half the lift is lost to test #1 being wrong.
SCENARIO C · THE WHOLE FORMULA FIRING
1.4×1.2×0.78=1.31×
Strong hypotheses, run in the order most aligned with how this audience decides.
— END OF DOSSIER 03 · ITERATIVE PRIORITIZATION
SECTION 07 / 08
Iterative prioritization · Common questions

Frequently asked questions.

5 questions

ICE, RICE, and impact-and-effort tell you what a test is worth and what it costs. They tell you almost nothing about whether it will actually win. Behavioural scoring runs every hypothesis through several psychological models (Fogg, Big Five, the seven drivers, brand-fit, historical priors) and produces a single composite priority — so the test at the top of the list is the one most aligned with how your buyers actually decide, not the one that sounds best in a planning meeting.

Four: Fogg Behaviour Model (Motivation × Ability × Prompt), Big Five personality traits (we mostly load on Conscientiousness and Neuroticism), our 7-driver framework (Security, Belonging, Mastery, Status, Novelty, Autonomy, Hedonism), and a fit-and-effort layer (brand alignment, build cost, historical priors from 4,000+ tests in Quantum). Each model produces a sub-score; weights combine them into one number.

Every test result — win, loss, or flat — feeds back into the weights. Over a quarter, the model stops being a static guess and starts being a calibrated prediction of what your specific audience responds to. We've seen win rates rise from ~38% in sprint 1 to ~84% by sprint 8 on the same backlog, just by re-weighting after each outcome.

Because priority is a function of who's buying. The same hypothesis can score 84 for a Security-seeker (evidence-first, risk-averse) and 57 for an Aspiration-driven buyer (identity, status, novelty). The interactive scoring card on this page is the literal interface — toggle the profile, watch the backlog rerank. Most teams pick one test for everyone; we pick the right test for the dominant profile in each segment.

It's the third variable in the growth formula: ΔRPU = Quality × Rate × Success rate. The Quantum Database covers Quality (4,127 indexed prior tests). Rapid A/B Testing covers Rate (up to 6 concurrent, multi-arm, audience-isolated). Iterative Prioritisation covers Success rate. Strong hypotheses without good prioritisation just means a faster roadmap of mediocre tests — the three only compound when all three are firing.

Iterative prioritization · Talk to us

Want your backlog rescored?

30-minute strategy call. We'll take your existing hypotheses, run them through the behavioural scoring model with your buyer profile, and send back the reranked queue plus the first three sprints we'd run.

  • €500M+ in additional revenue across 250+ brands
  • 4,000+ A/B tests · 52.6% win rate
  • 10% RPU uplift guaranteed in 6 months — or we work free
Book your free strategy callWe work exclusively with brands doing €300K+/month.
Drip Agency
About UsCareersResourcesBenchmarks
ImprintPrivacy Policy

Cookies

We use optional analytics and marketing cookies to improve performance and measure campaigns. Privacy Policy