Zac BrooksEvery business already runs a product loop. A transformation runs it on purpose, starting with the box that moves the most money.
- What this is
- A way to see a business as one loop, and a wheel of six dimensions, 32 areas and 103 capabilities to grade it on.
- Why it matters
- A transformation pays when its first change lands on revenue or cost. The wheel shows which box that is.
- How to use it
- Grade each area A to F, find the weak grade sitting next to the most revenue or cost, and move one number there first.
1The loop and the method
Every business runs the same loop, whether anyone has drawn it or not. It believes in something, takes it to market, delivers it and learns from what comes back.
Software teams have spent twenty years learning to run that loop on purpose. They change one thing, watch one number and keep what works. A transformation can run a whole business the same way.
1.1Business is a loop
Every turn of the loop compounds leverage or meets friction, and every box in it has a line to revenue or cost.
You believe in something, you take it to the market and try to sell it, you deliver what you sold, and the delivery gives you feedback, from the customer and against the OKR. Then you shape it. Done well, you compound leverage the whole way round.
Friction is whatever slows the turn: a slow handoff, a guess where a number should be, a step only one person can do. So the first move is never to burn it down and start again.
Map the business into small boxes and find where each one’s line to revenue and line to cost truly is. Sometimes the system is the problem. Often it is the strategy, and the boxes show which.
1.2Transformation runs like a product
The best SaaS teams ship features to shape how users behave, and the money follows across the whole product. AI-native transformation runs the same play on the business, aimed at one OKR that ties to revenue or cost.
As a business gets closer to AI-native, what it does starts to look like traditional product development. A product team and a business transformation run the same three steps, and only the middle one differs: a product team bets on what users will do, a transformation bets on an OKR.
- Map it. Grade every capability and find the red square.
- Bet on the gap. Write down what changing it will do.
- Track it. Watch a leading number move, then the money.
Cash-flow forecast is missing. The competitor has it.
Support is weak: a problem order takes three days to fix.
Why the target has to be an OKR
- No crowd to average over
A SaaS product lets a feature prove itself across thousands of users. A business has one P&L, so each change needs its own line to revenue or cost.
- Something you can grade
An OKR has a before and an after. The box gets a grade from A to F on whether it moved, with the evidence beside it.
- The order picks itself
The weak grade sitting next to the most revenue or cost goes first. That is the roadmap.
2Mapping the business
To run the loop on purpose, you need to see the business in pieces small enough to change. The wheel cuts it into 32 areas and 103 capabilities, and a grade on each one says where to start.
2.1The wheel
Six dimensions, each a question a CEO can answer about their own company. They fan out to 32 areas and 103 capabilities, the boxes we grade.
- Ring oneThe dimension, and the question it asks
- Ring twoIts areas: the boxes a business is mapped into
- Ring threeThe capabilities inside each area
Hub The business itself. People, systems and agents live across every dimension.
Can purpose, value, and problem become a thesis that sells?
GrowthDoes the market know about it, and will they pay?
ExecutionDoes delivery hold its profit and keep its customers as demand grows and the work changes?
InformationDoes what happens get seen, shared and learned from, without anyone stopping to record it?
IterationDoes each turn leave the business with more margin and more leverage?
RiskCan we do something incredible safely? Don’t lose trust through ignorance.
- 6dimensions
- 32areas
- 103capabilities
Drag to move. Hold Command or Control and scroll, or double-click, to zoom. Select any cell to read it.
2.2How a box gets graded
Each area gets a grade from A to F, with the evidence beside it. The grades say where to start.
- ARuns well, measured and improving
- BWorks, and is mostly measured
- CWorks on people’s effort
- DPainful, or run blind
- FMissing, or breaks under load
The wedge
Start where a weak grade sits next to the most revenue or cost.
The goal is a wedge first, not the whole wheel. It is usually something you already know: if the friction here dropped, something would change. Over time, working with the people who run the business, you capture one box, then the next.
3The six dimensions
Each dimension asks the business one question. Read them in order and the loop comes back: the thesis, the market, delivery, what gets seen, what gets better, and the floor that keeps it all safe.
3.1Mission
Can purpose, value, and problem become a thesis that sells?
- 6areas
- 20capabilities
Mission is the thesis: the problem you solve, for whom, and why now. It moves during a transformation, because as execution gets stronger the business can take on larger and more valuable problems.
The moat lives here too: what makes the position hard to copy.
Why the company exists, what it values, and why now.
- North star The one long-range goal every decision can be checked against.
- Values What the company will and will not do, written down and actually used to decide.
- Why now What changed in the market or the technology that makes this the moment.
Product and market research, and an honest read of what the business can do today. It feeds the thesis.
- Market maps Who is in the market, how big each segment is, and where the open ground is.
- Deep research Questions about customers, competitors and trends answered with sources, not guesses.
- Capability assessment An honest read of what the business can deliver today, and what it would take to do more.
The problems you solve, ranked by value and difficulty. Solved problems get more valuable as execution advances.
- Problem portfolio The problems you solve, listed with who has them and what solving each one is worth.
- Value & complexity ladder Problems ranked from simple and cheap to hard and valuable, so you know which to climb to next.
- Capability pairing Each problem matched to the capability that solves it, so the gaps are visible.
The niche you own, the category you claim, how you differ.
- Niche selection The specific customers you serve best, chosen on purpose.
- Category frame The category you tell the market you are in, and why it suits you.
- Competitive contrast How you differ from the alternatives, in words a buyer would repeat.
What stops a competitor from copying you: access they cannot get, a cost they cannot match, customers who would lose something by leaving, and know-how that took years to build.
- Exclusive access Supply, relationships, data or rights that competitors cannot get.
- Cost advantage You can deliver for less than a competitor could match, from your scale or from how you work.
- Switching costs Customers would lose something real by leaving: their history, their setup, or the way you fit how they work.
- Hard-won know-how How you do the work took years to learn, and cannot be copied by reading about it.
What the market thinks and feels when it hears your name: recognition, reputation, and a story and a look that hold together.
- Recognition People in your market know your name and what you do.
- Reputation What customers say about you when you are not in the room, backed by reviews, references and proof.
- Story A clear story of who you are, who you serve and why it matters, told the same way by everyone.
- Look & voice A look and a way of speaking that are consistent everywhere you show up.
3.2Growth
Does the market know about it, and will they pay?
- 6areas
- 20capabilities
Growth is every path from a stranger to a signed deal. Each path is a feature: you can count it, test it and improve it, the same way a product team treats a signup flow.
People find you: content and media, search, AI answers, the website and advertising.
- Content & media Articles, video, podcasts and posts that teach your buyers something, made on a schedule.
- Search (SEO) Buyers who search for the problem find you on the first page.
- AI answers (AEO) When buyers ask ChatGPT or another AI assistant about the problem, you are part of the answer.
- Website The site explains what you do in seconds and turns visits into conversations.
- Advertising Paid ads that reach the right buyers, judged on what they bring in rather than on clicks.
You find them: campaigns, signals on who needs you now, ranked leads.
- Campaign engine Outreach run as campaigns, each with a list, a message and a measured result.
- Demand intelligence Signals that show who needs you now, such as a new hire, new funding or a change in their market.
- Qualification & ranking Leads scored and ranked, so time goes to the ones most likely to buy.
Others bring you in: referrals, channels and co-marketing.
- Referral loops Happy customers and contacts send new business, and you make it easy for them.
- Channel alliances Partners who sell or bundle your offer to their own customers.
- Co-marketing Joint content and events with partners who share your buyers.
An audience you own and show up for.
- Owned audience A list or community you can reach without paying a platform.
- Live formats Events, webinars and sessions where buyers meet you in person or on screen.
- Social presence A steady presence on the channels your buyers use.
A problem, for a segment, married to a capability and priced as an offer. Often many, each tailored, tested in sales.
- Hyper-tailored packages Offers shaped for one segment’s problem, each with its own price and promise.
- Offer testing New offers tried with real buyers, and kept or cut on the result.
- Entry ladder A small first purchase that leads naturally to bigger ones.
Pipeline, proposals and pricing, and moving deals through.
- Pipeline ops Every deal tracked by stage, with a next step and a date.
- Proposals & pricing Proposals and prices that are fast to produce and consistent from deal to deal.
- Deal navigation Knowing who decides, what is in the way, and how to move each deal forward.
3.3Execution
Does delivery hold its profit and keep its customers as demand grows and the work changes?
- 4areas
- 12capabilities
Execution is the physics of delivery: how the work is laid out, where its limits are, and how it holds up today. It keeps the profit you priced and the customers you won as demand grows and the work changes.
Iteration is what moves those limits.
How the work is laid out: the steps, the handoffs, the playbooks, and who owns each part. It decides whether everything else in Execution is even possible.
- Playbooks The way each job is done, written down and actually followed.
- Handoffs Work passes between people, teams and agents without getting lost or redone.
- Clear owners Every part of the work has one owner who answers for it.
More volume and more variety without cost or quality slipping: room to grow, a set way to take on new kinds of requests without derailing the work, and a workforce that can carry the load.
- Volume headroom How much more work the current setup can take before cost or quality slips.
- Handling variety When new kinds of requests or problems come in, there is a set way to take them on, so they do not derail the work or the margin.
- Workforce capacity Enough people with the right skills to carry the load as it grows.
Is it accurate, and does it come out the same every time? Checks happen before anything reaches the customer.
- Accuracy The work is right the first time.
- Repeatability The same job comes out the same way every time, whoever does it.
- Quality checks Reviews and spot checks catch problems before the customer does.
When something goes wrong, it is caught, owned and fixed fast enough that the customer stays.
- Catching issues Problems are spotted early, often before the customer reports them.
- Time to fix When something breaks, it is owned and fixed quickly.
- Keeping the customer After a problem, the customer is looked after and stays.
3.4Information
Does what happens get seen, shared and learned from, without anyone stopping to record it?
- 6areas
- 21capabilities
Information is the business’s senses and nervous system. Software teams built this for code first: every change recorded, linked to its purpose and reviewed. With agents, the bottleneck moves from doing the work to absorbing it.
Everyone, people and agents, sees what is happening across the business in one place: who did what, which goal it served, what was said and decided, and where each customer and deal stands, not 50 tools that each hold a piece.
- Book of business Every customer, deal and order in one place, current and complete.
- Who did what A clear record of what was done and by whom, people and agents alike.
- Work tied to goals Each piece of work is linked to the goal or OKR it serves.
- Communications Emails, calls and messages are kept where anyone who needs them can find them, linked to the customer or the work.
- Ideas Ideas are captured where everyone can see and build on them, not lost in someone’s head or a private chat.
Doing the work records the work. Nobody spends hours a week updating tools so others can see what they did.
- Captured as you work Doing the work creates the record, with no separate step to log it.
- No status updates Nobody spends time reporting what they did; the record already shows it.
- Connected tools The tools the business uses feed one record instead of each holding a piece.
You can analyze what customers do, and what happens when you change something. Mostly numbers, and the more data there is, the more it pays.
- Customer behavior What customers buy, how often, and what makes them come back or leave.
- Change impact What actually happened when something changed, measured against before.
- Forecasts & anomalies What is likely to happen next, and an early warning when something looks off.
Work and knowledge pass quickly from one person or agent to another and are actually taken in. A new hire and a month of agent output both test it.
- Onboarding A new person gets up to speed fast, from the record rather than from someone’s memory.
- Absorbing agent work What agents produce is summarized and checked, so people can take it in.
- Context on demand Anyone, person or agent, can pull the background a task needs in moments.
The results of each change, what worked, what failed and why, come back into the record, so the next decision starts from them.
- Results written back The result of each change is recorded where the next decision will see it.
- Failure reviews When something fails, the cause is found and written down, without blame.
- Mining past work Past calls, emails and files are searched for what worked and what did not.
Out-of-date documents, stale playbooks, abandoned projects, duplicate tools and noise are cleared out, so what remains can be trusted. Agents make bloat grow faster.
- Stale information Documents and playbooks that are out of date are fixed or removed, so nobody acts on old instructions.
- Abandoned work Projects that stopped are closed out, not left half open.
- Duplicate tools Tools that overlap are retired, so each kind of information has one home.
- Noise Reports, channels and alerts that nobody reads are switched off.
3.5Iteration
Does each turn leave the business with more margin and more leverage?
- 5areas
- 15capabilities
Iteration is agile run on big data. What the analysis finds goes into one ranked queue, ships, gets measured, and the business keeps the margin or leverage it bought. Information ends at a finding, and Iteration starts there.
Feedback from customers, the OKRs and the data lands in one queue of changes, ranked by what each is worth and cut small enough to ship fast.
- Feedback intake Every finding, from a customer, an OKR or the data, arrives as a proposed change.
- Value ranking Changes are sized and ranked by how much revenue or cost they move.
- Small batches Changes are cut small enough to ship and learn from quickly.
How long one full cycle takes: from spotting something, to the change being live, to knowing whether it worked.
- Signal to change How fast a finding becomes a change that is live.
- Change to result How fast you know whether the change worked.
- Parallel runs Several changes tested at once, instead of one after another.
Every change carries a goal and a before-and-after, so you know whether it worked and keep it or roll it back.
- A goal per change Every change names the number it is meant to move.
- Before and after That number is measured before the change and after it.
- Keep or roll back Changes that work are kept, and changes that do not are undone.
Each turn widens the gap between price and cost: the work gets cheaper to deliver, the price follows the value, and less is spent fixing mistakes.
- Cost compression Each turn, the same work takes fewer steps, less time and cheaper inputs.
- Pricing Prices follow the value delivered, and rise as quality does.
- Less rework Fewer mistakes to fix, so less is spent fixing them.
Output per person keeps growing because repeatable work moves to systems and agents, custom work becomes product, and what is built once gets used again. Volume is where it shows.
- Automation Repeatable work moves to systems and agents, and people keep the judgment.
- Productization Custom work is turned into standard offers that sell and deliver the same way each time.
- Reuse What is built once, a template, a playbook or a dataset, gets used again instead of rebuilt.
3.6Risk
Can we do something incredible safely? Don’t lose trust through ignorance.
- 5areas
- 15capabilities
Risk is the floor under the rest. Moving fast with AI is safe when the limits and checks are in place before anything ships, and nothing the business needs rests on one person or one tool.
Contracts say what you promised, who owns the work and who carries the liability, including work an AI did, and the rules of your sector are met.
- Promises & liability Contracts state what you promised and who is liable, including for work an AI did.
- Ownership of the work It is clear who owns what you produce, including what AI produced from customer data.
- Sector rules The rules of your industry are known, and met.
Cash and runway are known, spend limits are set before money goes out, since AI and tool costs can climb overnight, and no single customer or bet is big enough to sink you.
- Cash & runway You know how much cash there is and how long it lasts.
- Spend limits Limits are set before money goes out, including AI and tool costs that can climb overnight.
- Concentration No single customer, supplier or bet is big enough to sink the business.
Customer data is protected, people and agents reach only what they need, and access ends when someone leaves.
- Data protection Customer data is stored safely and used only for what was agreed.
- Access control People and agents can reach only what their job needs.
- Access removal Access ends the day someone leaves or an agent is retired.
Approvals before an AI acts, limits on what it can touch, and a way to stop it.
- Approval gates A person approves before an AI takes an action that matters.
- Reach limits Each AI can touch only the systems and data it needs.
- Kill switches Any AI process can be stopped at once.
The key people and the tools, vendors and AI models the business could not run without, and the plan if one of them goes away.
- Key people Work that rests on one person is known, shared and backed up.
- Key tools & vendors The tools, vendors and AI models the business cannot run without are known, with alternatives ready.
- Fallback plans There is a plan for losing any one of them.
4The wheel, applied
The method is easiest to see run twice. First on a product, where good teams already work this way, then on a business, where the same steps aim at an OKR.
4.1A product team picks its feature
An invoicing app for small businesses grades itself against its strongest competitor and bets on its one red square.
Take an invoicing app for small businesses. Graded against its strongest competitor, most of its map is fine: invoices go out cleanly, payments land and the reminders mostly work. One square is red. Owners can’t see next month’s cash, and the competitor shows them on its home screen.
Map it
The team grades every capability against the competitor. The cash-flow forecast comes back an F, because the app has none.
ABACAABCACBBBCDCFCDBCCCDCABCCCash-flow forecast is missing. The competitor has it.
Bet on the gap
They write the bet down before they build: if owners can see next month’s cash, they open the app every week, and weekly owners are the ones who upgrade.
The feature A cash-flow forecastThe behavior Owners open the app every weekThe revenue More upgrade to the paid planShip it
A forecast on the home screen, drawn from the invoices already in the app. It takes six weeks, and the square is graded again on what shipped.
CBCDCash-flow forecast moves from F to B.
Track it
The team reads the behavior first: owners checking in every week. Upgrades follow a quarter later, and the bet is judged on both.
Owners checking in weekly 52% to 58%Upgrades to the paid plan Reads next quarter
The order of the readings matters. Behavior moves within weeks and money takes a quarter, so the team reads the behavior first and knows whether the bet is working long before revenue can say.
4.2A business picks its wedge
A parts distributor grades itself on the wheel and finds the weak grade sitting next to the most revenue.
Now a parts distributor. Graded on the wheel, most of its areas sit at a C and nothing is on fire. Support is the weak grade that matters. A problem order takes three days to fix, and the problems land on the customers who order every month.
Map it
Every area gets a grade. Support comes back a D, and it sits next to the biggest line of revenue in the business: repeat orders.
BCBCCBCDCDBCCCBDDFCDDCDDFCCBBCDCSupport is weak: a problem order takes three days to fix.
Bet on the gap
The bet names the change, the OKR it moves and the money behind it. Problem orders fixed inside a day keep the monthly customers ordering.
The change Problem orders fixed inside a dayThe OKR Repeat-order rateThe revenue or cost Repeat revenue keptChange the box
An agent reads each problem order, drafts the fix from the order history and routes it to a person for approval. The person who used to chase each fix now signs it off.
CCBBSupport moves from D to B, with a fix drafted the same day.
Track it
Time to fix moves first, from days to hours. The repeat-order rate follows over the next two quarters, and that is the number the business already runs on.
Time to fix a problem order 3 days to 5 hoursRepeat-order rate 61% to 64%Repeat revenue kept Reads next quarter
The shape is the product team’s, and only the bet differs. The product team bet on what owners would do in its app. The distributor bet on an OKR it already ran on, so the change shows up where the owner already looks.
4.3A discipline, and who it is for
Keeping the whole map graded is real work. It pays for a business investing deeply in AI, and a builder sells it one wedge at a time.
Running a business this way is a discipline. All 103 capabilities stay graded, the weak ones wait in one ranked queue, and every change is measured against the number it was meant to move. That is worth it for a business that means to invest deeply in AI, because each change compounds on the last. For everyone else the full map becomes a report in a drawer.
If you build for these businesses, know that the map overloads anyone who doesn’t think in products. Show an owner 103 boxes on a first call and they see an audit, and an audit with a fee up front is easy to refuse. Your job in the sale is to find the wedge without one.
The first conversation runs in four moves:
- Ask for the number. “If you could double one number in the business in ninety days, which would it be?”
- Walk back to the box. Find the one area that moves that number.
- Grade it together. Score that area in the room, with the owner, against the wheel.
- Price the wedge. One bet, its OKR and a date. The first win pays for the rest of the map.
The wheel stays in your pocket. The customer sees one wedge, one number and a date, and the rest of the map comes out once the first win has paid for it.