12The Aurygine Library
Operating knowledge
Thirty company-building decisions, written as systems rather than essays. Aury uses these inside your company's work — this page exists so you can see what it reasons with, and where each pattern stops being true.
Operating playbooks
30
Decision domains
17
Carry stated limits
30/30
- Market selection
Is this market worth being in?
The product works, customers say nice things, and the business still feels heavy.
- Smallest test
- Run ten qualified conversations in one adjacent segment with the current product and the current price. Change nothing else.
- Stop rule
- Fewer than two of ten reach a priced proposal — the adjacent segment is not better, it is just new
Sourced operator experienceAurygine synthesisConfidence: medium - Market selection
Should we pivot?
Growth has stalled and the founders disagree about what to do about it.
- Smallest test
- Put both directions in front of twenty real buyers in the same week, priced, and count which one produces more commitment.
- Stop rule
- Both directions produce almost nothing — the problem is upstream of the choice
Sourced operator experienceAurygine synthesisConfidence: medium - Customer and ICP
Who is this actually for?
Some customers are easy and profitable and others are painful, and nobody can say why.
- Smallest test
- Write the profile of the five best {unit}s as a one-page definition, then qualify the next ten inquiries against it and record the difference in {conversion}.
- Stop rule
- Matching and non-matching inquiries convert the same — the profile is describing a coincidence
Sourced operator experienceAurygine synthesisConfidence: high - Offer design
Designing the offer
People understand what the product does and still do not buy it.
- Smallest test
- Rewrite the offer as a single outcome with a fixed scope and a named first milestone, then use it on the next five proposals.
- Stop rule
- No change in {conversion} after ten proposals
Sourced operator experienceAurygine synthesisConfidence: medium - Pricing
Pricing as an experiment
The price was set early, nobody has changed it, and nobody knows if it is right.
- Smallest test
- Quote the next ten new opportunities at a higher price. Change nothing else, and do not touch existing customers.
- Stop rule
- Win rate collapses to a level that cannot fill {capacity}
Sourced operator experienceAurygine synthesisConfidence: medium - Commercial architecture
Monthly, annual or multi-year?
Revenue is real but cash is tight, and every customer pays monthly.
- Smallest test
- Offer one clearly-valued incentive for annual payment to the next five buyers, and record how many take it.
- Stop rule
- Nobody takes it at any reasonable incentive
Sourced operator experienceAurygine synthesisConfidence: medium - Commercial architecture
Separating what the customer pays from when the company gets paid
Customers can afford a monthly amount; the company needs the cash now.
- Smallest test
- Run one real transaction end to end under the new structure before offering it to anyone else.
- Stop rule
- The structure cannot be documented clearly
Sourced operator experienceAurygine synthesisConfidence: low - Acquisition
Designing the funnel backwards from capacity
Marketing is busy, the pipeline is thin, and nobody can say how many leads are actually needed.
- Smallest test
- Build the arithmetic chain once — target {revenue} → {unit}s needed → opportunities → leads → cost — and compare the required spend against available cash.
- Stop rule
- The chain cannot be completed because the data does not exist — collecting it becomes the mission instead
Sourced operator experienceAurygine synthesisConfidence: high - Sales
How much selling capacity does this actually need?
Deals are slipping and it is unclear whether the problem is leads, people or process.
- Smallest test
- For four weeks, protect a fixed block of selling time and count opportunities handled and closed within it.
- Stop rule
- The protected time cannot be defended — the real constraint is delivery, not sales
Sourced operator experienceAurygine synthesisConfidence: medium - Acquisition
What is an acceptable acquisition cost?
Paid channels are producing something, and nobody can say whether it is worth it.
- Smallest test
- Run one channel at a fixed, losable budget for four weeks with attribution recorded, and compute cost per {unit} and payback.
- Stop rule
- Cost per {unit} exceeds gross margin in the first year
Sourced operator experienceAurygine synthesisConfidence: medium - Sales
Founder-led selling, and when to stop
The founder is the only person who can close, and that is now the ceiling.
- Smallest test
- Write the call from records — questions, objections, answers, close — and have someone else run five conversations with it.
- Stop rule
- Every conversation needs the founder to rescue it — the offer is not ready to delegate
Sourced operator experienceAurygine synthesisConfidence: high - Retention
Getting a new customer to value
Customers buy and then go quiet.
- Smallest test
- Take the next five {unit}s through a written, owned two-week sequence and record time to first value against the previous five.
- Stop rule
- No change in time to value — the block is in the product, not the process
Sourced operator experienceAurygine synthesisConfidence: high - Retention
Why customers leave
Customers are leaving and the reasons given do not add up.
- Smallest test
- Classify the last ten departures into never-activated, wrong-fit, or value-decayed, using records rather than recollection.
- Stop rule
- Departures spread evenly across all three — there is no single cause and no single fix
Sourced operator experienceAurygine synthesisConfidence: high - Retention
Growing existing customers
New business is expensive and existing customers pay the same as on day one.
- Smallest test
- Offer one adjacent, already-requested capability to the five heaviest users at a stated price.
- Stop rule
- Nobody accepts
Sourced operator experienceAurygine synthesisConfidence: medium - Team and roles
Tying a role to an outcome it controls
People are busy, activity reports look fine, and the business is not moving.
- Smallest test
- Rewrite one role as a single controllable outcome with a measurement date, and review it after one month.
- Stop rule
- The outcome turns out to be outside the person's control — rewrite rather than enforce
Sourced operator experienceAurygine synthesisConfidence: medium - Team and roles
Paying for outcomes people control
A variable-pay scheme is rewarding the wrong behavior, or nobody trusts it.
- Smallest test
- Model the last two quarters under the proposed scheme and see who would have been paid what, before changing anything.
- Stop rule
- The model produces outcomes the company could not afford or defend
Sourced operator experienceAurygine synthesisConfidence: low - Team and roles
Making expectations explicit enough that the wrong hire self-selects out
Hires arrive, struggle with how the company works, and leave.
- Smallest test
- Rewrite the role honestly, including what people find hard, and compare the quality and fit of the next ten applications.
- Stop rule
- Applications fall to zero — the honesty has tipped into deterrence, or the role is genuinely unattractive and needs changing
Sourced operator experienceAurygine synthesisConfidence: medium - Product
Seeing what customers actually do
Roadmap decisions are being made from opinions because there is no usage data.
- Smallest test
- Instrument three actions and read two weeks of real usage.
- Stop rule
- Instrumentation would take longer than a week — reduce to one action
Sourced operator experienceAurygine synthesisConfidence: high - Product
Building what customers do, not only what they ask for
The request list is long, contradictory, and mostly from the loudest customers.
- Smallest test
- Interview five customers about what they do immediately before and after using the product, and count repeated behaviors.
- Stop rule
- No repeated behavior appears — do not build on a single account's request
Sourced operator experienceAurygine synthesisConfidence: medium - Product
Choosing integrations by the decision they unlock
There is pressure to add connectors, and no way to choose between them.
- Smallest test
- Rank candidate integrations by number of paying customers whose workflow it completes, and build only the top one.
- Stop rule
- Adoption below a tenth of requesters — requests were aspiration, not need
Sourced operator experienceAurygine synthesisConfidence: medium - Distribution
Being ready for a moment of attention
Something is about to send a lot of people at once — a launch, a feature, a broadcast.
- Smallest test
- Walk the full arrival path yourself as a stranger, end to end, and record every point where nothing happens.
- Stop rule
- The path cannot be completed before the event — reduce the event, not the readiness
Sourced operator experienceAurygine synthesisConfidence: medium - Distribution
Borrowing trust instead of buying attention
Paid channels are expensive and the company has no audience of its own.
- Smallest test
- Secure one appearance in front of a relevant audience, with a dedicated landing path, and measure arrivals and inquiries.
- Stop rule
- Two appearances produce nothing measurable — the message or the audience is wrong
Sourced operator experienceAurygine synthesisConfidence: medium - Partnerships
Partnerships that actually move revenue
Partnership conversations happen often and produce nothing.
- Smallest test
- Ask one partner for five warm introductions and measure {conversion} against the company's normal rate.
- Stop rule
- The partner does not deliver the five
Sourced operator experienceAurygine synthesisConfidence: medium - Experiments
The smallest meaningful test
A decision needs making and the proposed next step is large, slow or expensive.
- Smallest test
- Write the belief, the measure, the sample, the success rule and the stop rule on one page before doing anything.
- Stop rule
- The success rule cannot be written — the belief is too vague to test
Sourced operator experienceAurygine synthesisConfidence: high - Experiments
Scale, stop or change
A test has finished and the temptation is to interpret it generously.
- Smallest test
- Compare the recorded result against the recorded rule, in writing, before any argument about what it means.
- Stop rule
- No rule was set — the test cannot be interpreted; set rules and re-run
Aurygine synthesisConfidence: high - Capital
Is there anything worth raising against?
Raising is being considered, usually because cash is short or growth is wanted faster.
- Smallest test
- Write the use of funds as three milestones with dates and the evidence behind each, then show it to two informed outsiders before any investor.
- Stop rule
- The use of funds cannot be written concretely — the company is not ready to raise
Sourced operator experienceAurygine synthesisConfidence: low - Capital
Bootstrap, equity or debt
Growth is limited by cash and the funding route has not been chosen deliberately.
- Smallest test
- Model all three routes against the same plan, showing cash, cost, control and what happens if the plan underdelivers by half.
- Stop rule
- The plan the capital funds is not evidenced — no route is appropriate yet
Sourced operator experienceAurygine synthesisConfidence: low - Geography
Expanding into a new market properly
A new country or region is being considered, usually with more enthusiasm than evidence.
- Smallest test
- Serve three customers in the new market manually, at full price, and record the true cost to deliver.
- Stop rule
- Cost to serve destroys margin
Sourced operator experienceAurygine synthesisConfidence: low - Constraint diagnosis
What is actually holding this company back?
Something is wrong, effort is going in everywhere, and nothing is moving.
- Smallest test
- Fill in the five numbers from company records. The one that is worst relative to its own history is the candidate constraint.
- Stop rule
- Fewer than three of the five numbers exist — the mission becomes measurement, honestly stated
Sourced operator experienceAurygine synthesisConfidence: high - Company learning
Making the company's own evidence outrank generic advice
The same questions keep being re-answered from scratch.
- Smallest test
- After the next completed test, write the hypothesis, the result, the verdict and whether to repeat, and check it surfaces in the following decision.
- Stop rule
- Lessons are written but never influence a plan — the retrieval is broken, not the writing
Aurygine synthesisConfidence: high