100xLab

RES-06RESOURCES · PLAYBOOK

The GTM playbook.

How we take a product to market as a system instead of a scramble. Sixteen chapters, each with a play to run and a rule we don't break. Written in the lab, tested on our own launches before anyone else's.

16
CHAPTERS
5
PARTS
16
RULES WE KEEP
1
PERSON SIGNS OFF
UPDATED
2026-10-26

WHAT THIS IS

Go-to-market, engineered.

Most launches run on a date, a deck and a lot of optimism. That works about as often as optimism does.

Marketing engineering is the version that ships with instruments. Positioning written as rules. A site that explains itself in five seconds. Tracking in place before the first visitor. Content on a queue, not on someone's good week. Leads answered in minutes. And a person who signs off on everything that goes out.

It overlaps with growth, marketing ops and RevOps. The difference is that one team designs it, builds it and measures it as a single system, so nothing gets lost between departments.

00 · START HERE

Tick what sounds familiar.

Nobody reads a playbook front to back. Tell us what hurts and we'll tell you where to start.

Tick what sounds familiar.

READ FIRST

00 / 22

Nothing ticked. Either things are going very well, or you're being polite.

00.1 · THE JOB

What a marketing engineer actually builds.

Not campaigns. Machinery that keeps producing campaigns, with a person at the controls.

IN A WEEK, A MARKETING ENGINEER

  1. 01Writes the positioning as rules a sentence can pass or fail.
  2. 02Builds the site, pages and deck from those rules.
  3. 03Puts tracking in before traffic, and a dashboard nobody has to assemble.
  4. 04Structures facts so Google and AI assistants describe you correctly.
  5. 05Runs content from templates on a weekly queue.
  6. 06Routes every enquiry to the right person in minutes.
  7. 07Decides what gets automated, and what stays human.
Who does what
ROLEOWNSSHIPSMEASURES
Marketing engineerThe whole systemBrand, site, tracking, search, content, routingRevenue numbers, weekly
AgencyA retainerDeliverables per departmentActivity, monthly
Marketing opsThe toolsIntegrations and reportsData quality
FreelancerOne skillOne piece of the puzzleWhether the client is happy

00.2 · THE STACK

Six layers. Only one of them is clever.

Click a layer. The clever one is the rules file; everything else just has to work every day.

LAYER 01

Rules

Brand, voice, approved claims, banned words, gates and budgets.

Every person and every agent reads this first, so nobody improvises your company.

SEE CHAPTER 04

PART I

Define

Find the one bottleneck worth fixing, say who it's for, write down today's numbers and the rules.

CHAPTER 01 · DEFINE

Pick the bottleneck that costs the most

Most teams aren't short of people. They're short of a decision about what to fix first.

WHY IT MATTERS

Every growth problem looks urgent from the inside. If you automate the wrong one, you just get the wrong result faster. The first job is to find the single step where the most money leaks, and leave everything else alone for now.

THE PLAY

List the jobs your marketing does every week. Score each one on three things: how often it happens, how directly it touches revenue, and how clearly you could write down the rule a good person follows. Build only what scores high on all three. Then watch three real cases of it happening, a normal one, an awkward one and a failed one, before you design anything.

STEP BY STEP

  1. 01List every recurring job: replying to enquiries, publishing posts, building the report, updating pages.
  2. 02Score each 1 to 5 on frequency, revenue impact and how clearly the rule can be written.
  3. 03Throw out anything that's creative judgement every time. That stays human.
  4. 04Shadow three real cases of the winner and write down what people actually do, not what the process doc says.
  5. 05Write the bottleneck as one sentence: who is slowed down, at which step, and what it costs.

TRY IT · SCORE A BOTTLENECK

4

How often does this job happen?

4

How directly does it touch revenue?

3

Could you write down the rule a good person follows?

SCORE

48

A

Engineer it first

It happens often, it touches revenue and the rule fits on a card. Stop reading and start.

EXAMPLE · ILLUSTRATIVE

A clinic thinks it needs more content. Scoring says otherwise: booking requests arrive forty times a week, each one is revenue, and the rule (answer, offer two slots, confirm) fits on a card. Content can wait. Bookings go first.

COMMON MISTAKES

  • Fixing the job that's most annoying instead of the one that's most expensive.
  • Automating a task that only happens once a quarter.
  • Designing from the process document instead of watching real cases.
DONE WHEN

0 of 3 done

THE RULE

Fix one bottleneck at a time. The one that costs the most.

CHAPTER 02 · DEFINE

Position like a spec

If your positioning can't fail a test, it isn't positioning. It's a mood.

WHY IT MATTERS

Every system downstream reads your positioning: the site, the posts, the replies, the agents. If it's vague, every one of them improvises, and you end up with five slightly different companies.

THE PLAY

Write who you're for, who you're not for, the one problem you solve and three proofs. Phrase each as a filter a sentence can pass or fail. Then run your homepage, your deck and your last ten posts through it.

STEP BY STEP

  1. 01Who it's for: three filters you could check against a real account.
  2. 02Who it's not for: two filters, so the system knows when to say no.
  3. 03The one problem, in the buyer's words, taken from real calls.
  4. 04Three proofs you can link to: a result, a client, a demo.
  5. 05Run your current copy through the filters and cut what fails.

Paste a sentence to test it.

EXAMPLE · ILLUSTRATIVE

"We help businesses grow" fails every filter. "Engineering fabricators selling to industrial buyers who ask for specs before they ask for a call" passes, and tells the site exactly what to show first.

To see story-first positioning on screen, these product launch video examples (opens in a new tab) show how teams turn it into a 60-second video.

COMMON MISTAKES

  • Writing for everyone so nobody feels left out.
  • Proof that's really just an adjective.
  • Taking the problem from your pitch deck instead of your buyers.
DONE WHEN

0 of 3 done

THE RULE

If a sentence would fit a competitor's homepage, it's cut.

CHAPTER 03 · DEFINE

Measure day one

Everybody wants the after photo. Nobody takes the before.

WHY IT MATTERS

Without a baseline, every result is an argument. With one, it's a number. It also tells you whether the bottleneck you picked is as expensive as it looked.

THE PLAY

Before building anything, write down the five numbers this work will be judged on, where each one lives, today's value and who owns it. Include the time things take and the time they spend waiting, not just the totals. If a number can't be found, finding it is the first job.

STEP BY STEP

  1. 01Pick five numbers, each tied to the bottleneck.
  2. 02For each: the definition, the source and today's value.
  3. 03Split time spent working from time spent waiting.
  4. 04Agree how long you'll wait before judging the result.

TRY IT · BASELINE SHEET

NUMBER

TODAY

WHERE IT LIVES

Missing a value? Finding it is the first job.

EXAMPLE · ILLUSTRATIVE

Enquiries per week: 40. First reply: 9 hours on average, most of it waiting overnight. Booked calls: 11. Now everyone knows what better means.

COMMON MISTAKES

  • Measuring what's easy instead of what matters.
  • Changing the definition halfway through.
  • Judging after a week when the cycle is a month long.
DONE WHEN

0 of 3 done

THE RULE

No build starts without a baseline.

CHAPTER 04 · DEFINE

Write the rules file

An agent without a rules file is an intern with your logo and no supervisor.

WHY IT MATTERS

People and agents both need to know what they're allowed to say, who approves what, and when to stop and ask. Write it once and every system inherits it. Skip it and every system invents its own.

THE PLAY

One short file that every person and every agent reads before touching your brand: voice, approved claims with their proof, banned words, who signs off on what, spending limits and a named owner. Keep it short enough that people actually read it, and version it like code.

STEP BY STEP

  1. 01Voice in five bullets, with one example of right and wrong.
  2. 02Approved claims, each with its proof link.
  3. 03Banned words and claims, including anything legal flagged.
  4. 04The sign-off gate for each kind of output.
  5. 05Spending and volume limits, and the owner's name.
rules.md
# Rules file: [company]

## Voice
- Plain, short sentences. Sentence case.
- Numbers over adjectives. No hype words.
- [Three words that describe how we sound]

## Approved claims (each with its proof)
- [Claim] — [link to proof]

## Never say
- [Banned words, competitor names, unproven claims]

## Sign-off
- Pages, ads, outreach: [name] approves before anything ships
- Posts from the queue: [name] approves the week on Monday

## Limits
- Paid tools: [budget] per week, then stop and ask
- Outreach: [number] new contacts per day, per domain

## Owner
- [Name], reviewed every [month]

EXAMPLE · ILLUSTRATIVE

A new freelancer and a new agent both start on Monday. Both read the same file. By Wednesday, their drafts sound like the same company.

COMMON MISTAKES

  • A forty-page brand book nobody opens.
  • Claims with no proof attached.
  • No owner, so nobody updates it.
DONE WHEN

0 of 3 done

THE RULE

Nothing ships that breaks the rules file. Not even the good stuff.

PART II

Design the system

Spec it on one page, then build the brand, the site and the instruments it runs on.

CHAPTER 05 · DESIGN THE SYSTEM

Spec the trigger and the data

Ten minutes of spec saves ten days of rebuild.

WHY IT MATTERS

Most broken systems fail at the edges: the wrong trigger, missing data, an output that lands somewhere nobody looks. A one-page spec makes those edges visible before anything is built.

THE PLAY

Write four things on one page: what starts it, what data comes in, what data you add, and where the finished result goes. Then add the exception path: what happens when data is missing, wrong or late. Put the output where people already work, not in a new tool they'll forget to open.

STEP BY STEP

  1. 01The trigger: a form, a new record, a list, or a schedule.
  2. 02The inputs: exactly which fields arrive, and which can be empty.
  3. 03The additions: what you look up, and what you do when it isn't found.
  4. 04The destination: the inbox, channel or board people already use.
  5. 05The exception path: who gets the cases the system can't handle.

TRY IT · ONE-PAGE SPEC

WHAT STARTS IT
Trigger: A form is submitted
What comes in: (unknown)
What you add: (unknown)
Where it goes: (unknown)
When it can't: (unknown)

EXAMPLE · ILLUSTRATIVE

Trigger: the contact form. Inputs: name, work email, message. Added: company, size, which service they asked about. Destination: the founder's inbox with a three-line summary. Exception: personal email addresses go to a review list, not the bin.

COMMON MISTAKES

  • Guessing missing data instead of marking it unknown.
  • Sending results to a new dashboard nobody checks.
  • No plan for duplicates.
DONE WHEN

0 of 3 done

THE RULE

Unknown is an answer. A guess is a bug.

CHAPTER 06 · DESIGN THE SYSTEM

Brand as a system, not a PDF

A brand book in a shared drive is a museum exhibit. Nice to visit, never used.

WHY IT MATTERS

Every system you build will produce brand output at volume. If the brand only exists as a PDF, volume means drift. If it exists as tokens and templates, volume means consistency.

THE PLAY

Turn the brand into tokens and templates: colours, type, spacing, a page template, a post template, a deck template. Lock the parts that never change, and label the slots that do. When the system makes the right thing the easy thing, people stop redrawing your logo in a hurry.

STEP BY STEP

  1. 01Extract the tokens: colour, type, spacing, radius.
  2. 02Name the ten templates you use most.
  3. 03For each, mark the locked parts and the open slots.
  4. 04Put the templates where the team already works.

EXAMPLE · ILLUSTRATIVE

Weekly posts went from a designer and two days of back-and-forth to a template with three open slots. The designer now spends that time on the launch, where it counts.

COMMON MISTAKES

  • Templates so rigid nobody uses them.
  • Tokens that live only in a design file.
  • No owner for changes.
DONE WHEN

0 of 3 done

THE RULE

If it isn't a template, it'll be redrawn wrong by Friday.

CHAPTER 07 · DESIGN THE SYSTEM

The five-second site

You get five seconds. Most sites spend them on a slogan.

WHY IT MATTERS

Every channel sends people to the site. If the first screen doesn't explain what you do, who it's for and what to do next, every other system leaks at the last step.

THE PLAY

One page, one audience, one action. The first screen says what it is, who it's for and what to do next. Proof sits right under it. Every section earns its place by answering a question buyers actually ask, or it leaves.

STEP BY STEP

  1. 01Headline: what it is, in plain words.
  2. 02One line: who it's for and the result they get.
  3. 03One button: the next step, named by what happens.
  4. 04Proof directly under the fold: a client, a number, a demo.
  5. 05Test it on five people who've never seen it.

EXAMPLE · ILLUSTRATIVE

A job platform's first screen said "The future of work". It now says what it does, for whom, and the button reads "Post a job". Same traffic, a clearer next step.

COMMON MISTAKES

  • Two audiences on one page.
  • Three buttons that all say "Learn more".
  • Proof buried at the bottom.
DONE WHEN

0 of 3 done

THE RULE

One page, one audience, one action.

CHAPTER 08 · DESIGN THE SYSTEM

Instruments before traffic

Launching without tracking is flying without instruments. Exciting. Briefly.

WHY IT MATTERS

You can't fix what you can't see, and you can't see what you didn't measure from the start. Tracking added later misses the launch, which is the moment you most need to read.

THE PLAY

Write a tracking plan before the first page ships: every action that matters, its name, where it fires and which of your five numbers it feeds. Build one live dashboard from it. No monthly report that someone has to assemble by hand.

STEP BY STEP

  1. 01Name each event the same way everywhere.
  2. 02Map every event to one of your five numbers.
  3. 03Keep personal data out of every event.
  4. 04Build one dashboard with six charts, not sixty.
  5. 05Check the events fire before you announce anything.

EXAMPLE · ILLUSTRATIVE

On launch day the dashboard showed most sign-ups came from one comparison page nobody had promoted. By Friday it was in the main navigation.

COMMON MISTAKES

  • Tracking everything and reading nothing.
  • Event names that change every sprint.
  • Personal data in analytics.
DONE WHEN

0 of 3 done

THE RULE

No page ships without its event.

PART III

Build and reach

Get found, keep content coming, reach out like a person and answer like a machine.

CHAPTER 09 · BUILD AND REACH

Search, and the assistants that answer it

Your buyers ask Google. Then they ask an assistant. Both need the same facts, said the same way.

WHY IT MATTERS

More buying journeys now start with a question to an AI assistant. Assistants repeat what they can find and trust. If your facts are thin or inconsistent, someone else's page shapes the answer.

THE PLAY

Give search engines and AI assistants clear, structured, consistent facts: what you do, who for, where, how to start, who's behind it. Use the same wording on your site, your profiles and everywhere you're listed. Answer the questions buyers actually ask, on pages built to be quoted.

STEP BY STEP

  1. 01Ask the main assistants who's best in your category, and note what they say about you.
  2. 02List where they're wrong or silent, and which pages caused it.
  3. 03Write one facts page and keep every profile consistent with it.
  4. 04Answer the top ten buyer questions on your own site.
  5. 05Add structured data so machines can read it without guessing.
  6. 06Re-check the same questions every month.

EXAMPLE · ILLUSTRATIVE

An assistant described a fabricator as a trading company because an old directory listing said so. One corrected listing and a clear facts page later, the description matched the site.

COMMON MISTAKES

  • Chasing keywords nobody asks about.
  • Different descriptions on every profile.
  • Answers hidden in PDFs.
DONE WHEN

0 of 3 done

THE RULE

Say it the same way everywhere, or the assistants will say it their way.

CHAPTER 10 · BUILD AND REACH

The content engine

Inspiration is a lovely visitor and a terrible employee.

WHY IT MATTERS

Content that depends on one person's good week stops when that person gets busy. An engine keeps going, and frees the people for the ideas that matter.

THE PLAY

Build content on a queue: a weekly slot, templates for each format and one idea turned into many cuts. People choose the ideas, write the lines that matter and approve the queue. Systems resize, schedule and report back.

STEP BY STEP

  1. 01One idea a week, chosen by a person from real buyer questions.
  2. 02Cut it into a post, a short video, an email and a page section.
  3. 03Queue it, and have the owner approve the week in one sitting.
  4. 04Read the numbers back every Monday and keep what worked.

EXAMPLE · ILLUSTRATIVE

One customer question became a page answer, a sixty-second video and three posts. The page now ranks for the question, and the video gets shared by the sales team.

COMMON MISTAKES

  • Many ideas, one format, no consistency.
  • Publishing without reading the numbers.
  • Approving posts one at a time, all week.
DONE WHEN

0 of 3 done

THE RULE

One idea, many formats. Never many ideas, one format.

CHAPTER 11 · BUILD AND REACH

First lines are human

An agent can find ten reasons to write to someone. It shouldn't pick the words.

WHY IT MATTERS

Buyers can spot a template from across the inbox, and one careless message costs more trust than ten good ones earn. Research scales well. Judgement doesn't.

THE PLAY

Let systems gather the reasons to reach out: what the account just did, what it's hiring for, what it published, each with a date and a link. Verify the facts and the contact details before anything is written. A person writes or approves the opening line, and every reply reaches a person the same day.

STEP BY STEP

  1. 01Collect up to three dated, linkable reasons per account.
  2. 02Verify the contact and the facts, and drop anything unverified.
  3. 03Rank reasons by how close they sit to the one problem you solve.
  4. 04A person writes or approves the first line.
  5. 05Send at a pace that protects your name, and route every reply to a person.

EXAMPLE · ILLUSTRATIVE

A short list of twenty accounts, each with a real reason and a human first line, booked more calls than the five-hundred-contact sequence it replaced.

COMMON MISTAKES

  • Research without verification.
  • Volume as a strategy.
  • Replies left for the system to handle.
DONE WHEN

0 of 3 done

THE RULE

An agent can draft. It never sends unread.

CHAPTER 12 · BUILD AND REACH

Answer in minutes

The fastest reply usually wins. So make speed the system's job, not someone's mood.

WHY IT MATTERS

Interest cools by the hour. An enquiry answered in minutes, with context, feels like attention. One answered tomorrow feels like a queue.

THE PLAY

When an enquiry lands, look up the company from the email domain, check it against your filters, route it to the right person and send an honest first reply, day or night. The person follows up with the summary already on screen. Anything the system can't place goes to a review list, never into silence.

STEP BY STEP

  1. 01Look up the company from the email domain.
  2. 02Check it against your for and not-for filters.
  3. 03Route by segment to a named owner.
  4. 04Send a first reply that answers their question directly.
  5. 05Send unclear cases to a review list with a daily check.

EXAMPLE · ILLUSTRATIVE

Enquiries that arrived overnight used to wait until lunch. Now each one gets a useful reply in minutes and lands in the right inbox with a three-line brief.

COMMON MISTAKES

  • Auto-replies that say nothing.
  • Routing rules nobody owns.
  • Dropping enquiries the system can't classify.
DONE WHEN

0 of 3 done

THE RULE

No enquiry waits longer than your reply promise.

PART IV

Prove it

Run it by hand, test it small, compare it in the shadows, then let it loose on a slice.

CHAPTER 13 · PROVE IT

Run it by hand, then on a small batch

Automating a bad decision just makes the same mistake faster.

WHY IT MATTERS

Most problems show up in the first few dozen cases: missing data, odd formats, a research step that sounds right and isn't. Finding them on a small batch costs minutes. Finding them on thousands costs your reputation.

THE PLAY

Before automating, run five to ten real cases by hand, including a messy one and one that should be excluded. Write down what was accepted, what was rejected and how long it took. Then run the system on a small batch, twenty-five to fifty items, and read every single output.

STEP BY STEP

  1. 01Run five to ten cases manually, including a messy one.
  2. 02Record acceptance, time and what people said.
  3. 03Run the system on twenty-five to fifty items.
  4. 04Read every output. Check the facts, the routing and the tone.
  5. 05Fix, then run the batch again before going further.

EXAMPLE · ILLUSTRATIVE

The research step looked perfect for the first ten accounts, then started inventing hiring news on the thirtieth. Caught on a batch of fifty, fixed in an afternoon.

COMMON MISTAKES

  • Skipping the manual run because the idea seems obvious.
  • Reading the first ten outputs and assuming the rest.
  • Fixing at scale what should have been fixed small.
DONE WHEN

0 of 3 done

THE RULE

Read every output in the first batch. Every one.

CHAPTER 14 · PROVE IT

Roll out in four steps

Autonomy is a privilege. Systems earn it the same way people do.

WHY IT MATTERS

A system that works in testing can still fail with real people, real volume and real timing. A staged rollout limits the damage while you learn, and builds the trust that lets you hand over more.

THE PLAY

Run it in the shadows first, beside the people, and compare its decisions with theirs. Then give it a small slice with clear limits and a named person watching. Then let it handle everything behind a person's approval. Only then go hands-off, and only for the patterns that held up.

STEP BY STEP

  1. 01Shadow: it runs, nothing it makes is used, you compare.
  2. 02Small slice: one segment, volume and cost caps, a named watcher.
  3. 03Human gate: it handles everything, a person approves each output.
  4. 04Hands-off: only proven patterns, spot-checked weekly.
  5. 05Write down the stop conditions before each step.

THE LADDER

STEP 01 · SHADOW

It runs next to the human process. Nothing it makes is used. You compare.

EXAMPLE · ILLUSTRATIVE

In shadow mode the routing disagreed with the team on one case in five. Every disagreement traced back to one missing rule. With it added, the slice went live a week later.

COMMON MISTAKES

  • Going hands-off on day one.
  • No stop conditions, so nobody knows when to pull the plug.
  • Treating every disagreement as the system's fault.
DONE WHEN

0 of 3 done

THE RULE

Autonomy is earned, one step at a time.

PART V

Operate

Measure it, decide to scale or cut, and hand it to someone who can run it without you.

CHAPTER 15 · OPERATE

Measure, then scale or cut

One dashboard. Three questions. Is it running, is it working, is it worth it.

WHY IT MATTERS

Sunk cost keeps bad systems alive. Clear numbers let you scale what works and switch off what doesn't, without a three-week debate.

THE PLAY

Tag every page, post, email and lead with the same fields so results can be traced back to the work. Measure in three tiers: health, performance and efficiency. Compare against the baseline from chapter 3. Scale what moved the number, fix or cut what didn't, and say plainly when the effect is still uncertain.

STEP BY STEP

  1. 01Tag everything: source, campaign, version, owner and approval status.
  2. 02Health: is everything running, with no silent failures?
  3. 03Performance: is it moving the five numbers?
  4. 04Efficiency: what does each result cost, in money and hours?
  5. 05Decide: scale, fix, or cut. Write the decision down.
The metric stack
TIERQUESTIONEXAMPLES
HealthIs it running?Events firing, forms delivering, queue publishing, routing working, no silent errors
PerformanceIs it working?Visitors to sign-ups, enquiries per week, time to first reply, assistants naming you
EfficiencyIs it worth it?Cost per enquiry, hours saved per week, cost of tools per result

TRY IT · SCALE OR CUT

CHANGE+30%

Scale it.It moved the number by more than a tenth. Put it on the next segment or service.

EXAMPLE · ILLUSTRATIVE

Two systems went live together. One cut reply time in half. The other produced reports nobody read. The first got scaled to a second service, the second got switched off.

COMMON MISTAKES

  • Keeping a system because of what it cost to build.
  • Calling a faster step a win when the total didn't move.
  • Reporting activity instead of outcomes.
DONE WHEN

0 of 3 done

THE RULE

Every automation has a health check and an owner.

CHAPTER 16 · OPERATE

Hand it over, then pick the next bottleneck

A system that only works when its builder is in the room isn't a system. It's a hostage situation.

WHY IT MATTERS

The point is marketing that keeps working after the build. That only happens when someone else can run it, fix it and pause it, without calling the person who built it.

THE PLAY

Turn recurring exceptions into written rules. Write a short runbook: how to run it, how to pause it, how to recover from a failure, where the rules live. Have the new owner run a normal case, diagnose a failure and pause it, while you watch. Then go back to chapter 1. The bottleneck has moved.

STEP BY STEP

  1. 01Turn recurring exceptions into written rules.
  2. 02Write the runbook: run, pause, recover, change.
  3. 03Move every account and access to the client's name.
  4. 04The new owner runs a handover test while you watch.
  5. 05Note what's reusable, then pick the next bottleneck.

EXAMPLE · ILLUSTRATIVE

The handover test found one step only the builder knew: a weekly manual refresh. It went into the runbook, then into the schedule. The owner hasn't needed to call since.

COMMON MISTAKES

  • Accounts left in the builder's name.
  • A runbook nobody has tested.
  • Stopping after one system instead of moving to the next bottleneck.
DONE WHEN

0 of 3 done

THE RULE

If only the builder can run it, it isn't finished.

APPENDIX

The appendix

A · AGENT SKILLS

Skills for the agents doing the prep.

Instruction files that teach an agent a repeatable marketing job: audits, briefs, launch plans and more. Pick one per chapter and let it do the prep.

B · THE OPTIONS

Agency, freelancers, a hire, or a lab.

Ways to run go-to-market
OPTIONGOOD FORWATCH FOR
AgencyBig campaigns with a big budgetHand-offs between departments, billing by the hour, reports about activity
FreelancersOne skill, quicklyYou become the project manager for five people who never meet
A hireLong-term ownership in-houseMonths to hire, one person's range, cost rises with every new role
100xLabThe whole system, built and measured by one senior teamWe build one system at a time. If you want volume for its own sake, we're the wrong call

C · THE WORK LOG

Where these plays came from.

Every chapter was run on our own launches first, then on client work. Here are the live sites, with what we worked on for each.

D · GLOSSARY

The words we use.

Marketing engineering
Marketing built like a product: designed, built, measured and improved as one system.
GTM
Go-to-market: how a product reaches the people who should buy it.
Positioning filters
Who you're for and not for, written so a sentence can pass or fail.
Rules file
The one document every person and agent reads before touching your brand.
Baseline
Today's value of the numbers you'll judge the work on.
GEO
Generative engine optimisation: clear, structured facts so AI assistants describe you correctly.
Tracking plan
The list of actions you measure, what each is called and which number it feeds.
Moment
A dated, sourced event that makes your message more relevant.
Shadow mode
An automation running beside the human process, with its output compared and not used.
Sign-off gate
The point where a person approves an output before it ships.

QUESTIONS

Questions, answered.

01What is a GTM playbook?

A GTM (go-to-market) playbook is a written plan for taking a product to market: who it is for, what to say, where to say it and how to measure it. This one is written as plays and rules, so a team or an agent can run it step by step.

02Is this just GTM engineering with a new name?

Close relatives. GTM engineering usually means the outbound machine: lists, data and sequences. Marketing engineering covers the whole system a buyer meets: brand, site, tracking, search, content, routing and outreach, built by one team.

03Do I need to code to use this?

No. Each play is written for people first. Where an agent helps, you need someone who can judge its output, which is the part no tool does for you.

04Where should I start?

Tick the symptoms at the top of the page. If you'd rather not, start with chapter 1: pick the one bottleneck that costs the most. Everything else follows from it.

05Will agents run my marketing?

They'll run the repeat work. People choose the ideas, write the lines that matter and approve what ships. Chapter 14 is how that trust gets earned, one step at a time.

06How long does it take to put a system in place?

With us, a Sprint puts one system live in a week, at a fixed price. Doing it yourself depends on your team's time; the chapters are ordered so each one pays off on its own.

07Can you just do it for us?

Yes. Book a 30-minute call and we'll tell you which system to build first. We reply within 4 working hours (IST).

NEXT STEP

Run chapter 1 this week.

Or hand us the whole playbook. One 30-minute call, then a fixed quote within 24 hours.