The Growth Stack
I map seven connected layers so each new campaign, workflow and tool has a clear job.
My Growth Stack framework maps the seven capabilities I look for in a working growth system: tracking, CRM, automation, acquisition, conversion, retention and reporting. I use it to check ownership and handoffs, identify the weakest dependency and choose the next practical improvement before adding software or increasing spend.
Build on a foundation you can trust
Read the layers from bottom to top. This is a map of capabilities; actual customer journeys and feedback travel in both directions.
The connection is part of the layer
I do not count a form as working just because it looks correct. I check whether an enquiry reaches the right CRM record, keeps its source, gets an owner and receives the promised response.
The same test applies after a sale. If delivery cannot see what was promised, the retention layer has a handoff gap even if the CRM shows a win.
The order does not mean I build acquisition before making an offer page. I establish a complete minimum journey, then strengthen the dependencies around it. The AI Revenue System explains how capture, follow-up, conversion and retention fit into that journey.
I use the self-assessment below to turn this map into a short list of repairs.
Define the minimum useful system
These tool names illustrate categories. I check current capabilities, cost, permissions and integration fit for the actual business before choosing a product.
1. Data and tracking
- Purpose
- I establish what happened, where demand came from and which outcomes can be trusted. A shared event definition is more useful than several dashboards counting different things.
- Minimum viable version
- One documented event map, consistent campaign tags, a unique lead identifier and a checked path from a test enquiry to a recorded sale. Collect only the data needed and apply appropriate consent controls.
- Signs it is broken
- Duplicate conversions, missing campaign sources, unexplained differences between forms and CRM records, or a dashboard that cannot trace a sale back to its enquiry.
- Tool categories and examples
- Web analytics such as GA4; tag management such as Google Tag Manager; consent management; call tracking; a spreadsheet or warehouse for reconciliation. These are category examples, not a required shopping list.
- Connection to the system
- Tracking gives CRM records an origin and gives reporting a consistent vocabulary. CRM outcomes feed back into acquisition analysis without placing personal details in analytics events.
2. CRM and pipeline
- Purpose
- I give each enquiry a record, a stage, an owner and a next action. The CRM becomes the working history that sales and operations can both use.
- Minimum viable version
- One pipeline with written entry and exit rules, deduplication, an accountable owner and a dated next step for every active opportunity. Keep contact permissions alongside the record.
- Signs it is broken
- Unassigned leads, personal spreadsheets replacing shared records, stages with no exit criteria, missing loss reasons, or deals whose next action nobody can explain.
- Tool categories and examples
- CRM and pipeline platforms such as GoHighLevel, HubSpot or Pipedrive; contact management; task queues; scheduling. I choose around the process, access needs and reporting requirements.
- Connection to the system
- It receives source identifiers from tracking and outcomes from conversion assets. It supplies automation with trustworthy stage changes and hands won customers to retention.
3. Automation and AI agents
- Purpose
- I turn repeatable handoffs into monitored workflows. Voice, chat and WhatsApp agents can answer routine questions or qualify an enquiry when the scope and permissions are clear.
- Minimum viable version
- One tested routing workflow, one follow-up sequence with reply and opt-out stop rules, an error queue and a human handoff. An agent gets approved information and a narrow task before it gets wider access.
- Signs it is broken
- Duplicate messages, sequences continuing after a booking, silent integration failures, invented agent answers, or customers unable to reach a person.
- Tool categories and examples
- Workflow builders such as GoHighLevel workflows, HubSpot workflows, Make or n8n; voice and chat platforms; approved WhatsApp Business integrations. I check channel policies, permissions and data handling separately.
- Connection to the system
- CRM state triggers the work; replies and bookings update CRM state. Agents use approved offer information from conversion assets and pass exceptions to a person rather than looping indefinitely.
4. Acquisition
- Purpose
- I bring relevant demand into the system through paid campaigns, SEO and answer engine optimisation, content or partnerships. I judge the source by downstream fit and economics.
- Minimum viable version
- One defined audience, one offer and one channel with tagged links, an accountable budget or time allocation, and a way to connect enquiries to qualified opportunities.
- Signs it is broken
- More clicks with no qualified pipeline, unknown source quality, conflicting campaign promises, or channel reports that end at cost per lead.
- Tool categories and examples
- Ad platforms such as Google Ads or Meta Ads; Google Search Console; SEO research and content publishing tools; partner referral tracking. I add channels when the current path can be measured and handled.
- Connection to the system
- Campaigns send demand to conversion assets and preserve source data for the CRM. Reporting returns qualified pipeline and customer outcomes so channel decisions reflect more than traffic.
5. Conversion assets
- Purpose
- I help a suitable buyer understand the offer and take a useful next step. Offers, landing pages, forms and funnels should make the promise and the handoff clear.
- Minimum viable version
- One clear offer page, an accessible labelled form or booking route, a meaningful confirmation and a tested CRM handoff. Check the journey on mobile and with a keyboard.
- Signs it is broken
- A message mismatch between ad and page, forms that fail silently, unnecessary required fields, mobile friction, or bookings that never appear in the pipeline.
- Tool categories and examples
- Landing-page and funnel builders such as GoHighLevel or HubSpot; WordPress or a custom website; form and scheduling tools; controlled experiment tools when volume supports them.
- Connection to the system
- Assets fulfil acquisition promises, send the source and enquiry to the CRM, and trigger the right automation. Results return to tracking so a successful form is distinguished from a page visit.
6. Retention and expansion
- Purpose
- I connect the sale to onboarding, delivery, renewal and appropriate expansion. A growth system must recognise customer outcomes after the initial close.
- Minimum viable version
- A named onboarding owner, a delivery milestone, a renewal or check-in date where relevant, and a place to record cancellation reasons and communication preferences.
- Signs it is broken
- Customers repeat their requirements after the sale, nobody owns onboarding, renewals arrive as surprises, or acquisition looks healthy while customers leave.
- Tool categories and examples
- CRM customer pipelines; customer success and help-desk systems such as HubSpot Service Hub or Zendesk; billing and renewal tools such as Stripe where suitable; task and survey tools.
- Connection to the system
- A won CRM deal initiates delivery with context. Customer feedback improves the offer and acquisition targeting; renewal and expansion outcomes feed reporting. A completed support request is not automatic marketing permission.
7. Reporting and decision cadence
- Purpose
- I turn the other layers into a repeatable decision process. The report needs to explain what changed, where confidence is limited and who acts next.
- Minimum viable version
- One reconciled view of spend or effort, qualified pipeline, wins and customer outcomes, plus a weekly review with a decision log, an owner and a check date for each action.
- Signs it is broken
- Multiple definitions of revenue, dashboards with no decisions, unowned action items, or repeated changes before enough evidence has arrived.
- Tool categories and examples
- A spreadsheet, native CRM reports, Looker Studio or a business intelligence tool such as Power BI. I start with a report people can reconcile and use before adding more presentation layers.
- Connection to the system
- Reporting consumes evidence from every layer and sends priorities back down the stack. It should reveal a tracking gap or handoff failure instead of treating all disappointing outcomes as an advertising problem.
Score the evidence you have
I use this as a planning conversation, not a performance prediction. Choose a score for all seven layers; no answers are preselected.
0 - Missing or unreliable. I cannot demonstrate the minimum working path.
1 - Present with gaps. I can demonstrate parts of it, but ownership, handoffs or checks are inconsistent.
2 - Owned, connected and verified. I have a named owner and recent evidence that the minimum path and its handoffs work.
Choose a score for each layer, then calculate your total out of 14.
Read the score with its weakest layer
0-5: establish the foundation. Demonstrate one complete enquiry-to-outcome path before expanding the stack. 6-10: connect and verify. Repair ownership gaps and handoffs, then repeat the journey test. 11-14: refine the constraint. Use operating evidence to choose a focused improvement and monitor its effect.
These bands are my suggested planning rubric, not research findings, a maturity certification or a comparison with other companies. The total is a simple sum with equal weight per layer. A high total can hide a critical zero; start with missing dependencies and inspect tied low scores together.
Turn the map into a working cadence
1. Trace one real journey
I choose an offer and follow a permitted test enquiry from source to sale and onboarding. I record the systems involved, the handoff evidence and the owner of each exception.
2. Score with evidence
I ask each layer owner to demonstrate the minimum path. Missing evidence stays a gap. I score again after major changes, rather than treating the first assessment as a permanent rating.
3. Repair one dependency
I choose a small change with an owner and a check date, such as preserving source data in the CRM or stopping reminders after a booking. I test failure handling as well as the successful path.
4. Review outcomes weekly
I reconcile source data, qualified opportunities, wins and customer outcomes. I review failed workflows and customer friction, then decide whether to keep, change or reverse the repair.
A hypothetical priority decision
Suppose a team scores tracking 0, CRM 1, automation 2, acquisition 2, conversion 1, retention 1 and reporting 1. The total is 8 out of 14. I would first investigate the missing tracking path and CRM ownership before treating the automation score as dependable or increasing campaign spend. These are example scores, not client data.
For ongoing ownership I offer fractional growth support. For a specific workflow I use AI automation services, and for channel execution I connect the measurement plan to performance marketing.
Using the Growth Stack
What is the Growth Stack framework?
It is my seven-layer map of the capabilities a growth system needs, from data and tracking through CRM, automation, acquisition, conversion, retention and reporting. I use it to find missing capabilities and broken handoffs before adding more software.
Do I need a different tool for every layer?
No. One platform may cover several layers, and a spreadsheet can be enough for an early reporting process. I score ownership, working connections and checked evidence rather than the number of subscriptions.
Why are the layers shown in this order?
The diagram is a capability map, not a chronological customer journey or a strict deployment schedule. Data and CRM support dependable handoffs, while reporting reviews the whole system. A buyer often moves from acquisition to a conversion asset before entering the CRM.
How is the self-assessment calculated?
Give each layer zero for missing or unreliable, one for present with important gaps, or two for owned, connected and verified. The seven scores add to a total out of fourteen. The bands are my planning guide, not an externally validated benchmark.
Does the assessment store or send my answers?
No. The assessment totals the scores in your browser and does not save them, send them to me or make network requests. The site separately loads analytics and other shared services as described in the Privacy Policy.
Which layer should I fix first?
I start with the earliest dependency blocking a reliable customer handoff. A zero in tracking or CRM can make later scores difficult to trust. I then choose one small repair, assign an owner and repeat the journey test before increasing spend or automation.