The manifesto

Built around how people really think.

This page is the design standard we hold ourselves to, the rubric we hire against, and the document we hand a client in week one so nobody has to guess how a decision will be made. Five principles, each with a working demonstration you can touch.

Every Merge5 case study is tagged with two or three of these. When a feature cannot be argued against one of them, it does not ship.

013 taps to a confirmed time in WhenWouldWork

Zero-Friction Entry

Every screen before the first moment of value is a tax the user pays for our indecision. We design backwards from the aha and delete everything in front of it.

  • No account wall before the first win. Accounts come after proof.
  • One primary action per screen, sized for a thumb in motion.
  • Cold-start content so the app is never empty on day one.

Dopamine prediction error: the first reward must arrive faster than the user's doubt.

Time to first value54s

Fifty-four seconds from open to a real win. No account, no tour, no empty state.

0268% fewer coordination threads in WhenWouldWork groups

Single-Source Clarity

Duplicated state is duplicated doubt. When a product holds one canonical version of every fact, the interface gets quieter and the team moves faster.

  • One record, one owner, many views. Never two editable copies.
  • Interface copy mirrors the data model, not the org chart.
  • If two screens disagree, the bug is architecture, not styling.

Working memory holds roughly four chunks. Every duplicate consumes one.

Clutter slider4 sources of truth
Duplicate record
Second source of truth
Derived cache
Orphan field
Four copiesOne truth
034 questions in ReflectFour, 40 fields available behind them

Progressive Disclosure

Power should be discovered, not survived. We let the first session stay simple, then reveal depth to the people who ask for it.

  • Advanced controls live one deliberate gesture away.
  • Defaults are opinions, and ours are opinionated enough to ship.
  • Onboarding teaches by doing, never by explaining.

Cognitive load is a budget. Spend it on the decision, not the interface.

Progressive disclosureDepth 1 / 4

The one action

What the user came to do. One control, sized for a thumb, no navigation in the way.

0441% week-four retention across the Merge5 habit apps

Reward on Schedule

Habit is a schedule of rewards, not a streak counter. We design the loop so the reward is genuine progress, then we time it so it lands.

  • Variable reward, predictable value: surprise the timing, never the substance.
  • Celebrate completion, not compulsion. No dark-pattern streaks.
  • Every notification earns its interruption with something new.

Variable-interval reinforcement creates the strongest, most durable repetition.

Reward schedule
week 1week 4week 8

Unpredictable timing, genuine value. The loop that survives week four.

059 apps shipped, 5 weeks each, zero relaunches

Ship in Five

Scope is a design material. Five weeks forces the decision that a nine-month roadmap lets you postpone forever, and it puts real users in front of the work while it is still changeable.

  • Week 1 decisions, week 2 prototype, week 3 build, week 4 harden, week 5 ship.
  • A demo every Friday, in the real product, on a real device.
  • Anything that cannot ship in five weeks becomes the next sprint's problem.

Short feedback loops convert uncertainty into learning before motivation decays.

Delivery cadenceWeek 1

Scope, principles and the one job. Everything that cannot ship in five weeks is cut here, in writing.

Friday demo · 1 of 5 complete

Section A

Cognitive psychology & neuroscience

Six findings that decide more interface questions than taste ever will. We model the mental budget first, then design the screen that fits inside it.

Attention

Working memory holds about four chunks, and switching costs a measurable fraction of a second each time.

One primary action per screen. Navigation never competes with the decision.

Memory

Spaced retrieval beats rereading by a wide margin, but only when the interval is tuned to forgetting.

Sessions schedule themselves around what you got wrong, not what you have open.

Motivation

Dopamine tracks prediction error, not reward size. Surprise sustains a loop that a fixed prize cannot.

Variable timing, constant value. No streak guilt, no manufactured loss.

Emotion

People remember the peak and the ending of an experience, not the average of it.

Every flow is designed to end on a moment worth remembering.

Effort

Perceived effort rises faster than actual effort. A visible step feels twice as heavy as an invisible one.

Progress is shown as distance covered, never as distance remaining.

Trust

Trust is built by predictability. Inconsistent state reads as unreliability, not as a bug.

One source of truth per fact, everywhere it appears.

Section B

Product philosophy

01

One job, done relentlessly

A product earns its place by doing one thing better than the alternative. We name that job in week one and refuse every feature that does not serve it.

02

Opinionated defaults

Settings are where products go to die. We ship a point of view, then let people change it once they have a reason to.

03

Delete before you add

The last week of every sprint is spent removing. What survives deletion is the product.

Section C

Delivery cadence

Five weeks, five Fridays, one store build. Cadence is what makes quality repeatable instead of lucky: the same five decisions, in the same order, every sprint.

Select a week to see what happens inside it. The clock starts the Monday after kickoff, not the day a contract is signed.

Delivery cadenceWeek 1

Scope, principles and the one job. Everything that cannot ship in five weeks is cut here, in writing.

Friday demo · 1 of 5 complete

Section D

Engagement workflow

01

Kickoff

One call, ninety minutes. Scope, constraints and the single job written down before anyone opens an editor.

02

Shared board

One board, visible to you, with every decision dated. No status reports about the work, only the work.

03

Friday demo

A live demo in the real product on a real device, every week, at the same time. Recording if you cannot make it.

04

Handover

Repository, store accounts and a written decision log. You own everything, including the reasoning.

If this is how you want to build, week one is one form away.

Start a project inquiry