Measurable growth,
in the AI-search era

Growth is an engineering discipline, not a copy exercise. We wire the tracking, build the attribution, run the experiments and prepare the site to be cited by the generative engines that now sit between searcher and link.

See six capabilities Scope a growth build
Event stream flowing from site into a dashboard, with AI-search prism on the side growth.metrics cac · ltv · aio tracking · attribution · seo · geo Site · app page · view_item add_to_cart · purchase Warehouse 7-day events · +28% CAC $38.40 Δ −12% · 30d LTV $812 LTV/CAC 21.1 Paid search roas 4.8× AI-search prism entities · citations · feeds schema.org structured QA ai-bots allow A/B test +6.2% conv · 95% CI n=42,118

CAPABILITIES

Six disciplines under one growth contract

Tracking, attribution, CRO, technical SEO, AI-search readiness and executive reporting. Run together, not assigned to six agencies.

01

Tracking architecture

First-party event schema, consent-mode wiring, server-side tagging, durable storage. Built to survive the third-party cookie wind-down.

02

Attribution

Multi-touch, data-driven, MTA with ML enrichment. Marketing spend mapped to actual revenue, not just last click.

03

Conversion optimisation

Experiment platform, guard metrics, sequential testing, cohort view. Wins that survive a three-month cohort aren't flukes.

04

SEO · technical

Core Web Vitals, rendering strategy, crawl budget, structured data, internal linking, hreflang.

05

GEO · AI-search

Readiness for generative search (ChatGPT, Perplexity, Google AIO): entity clarity, citation-worthy content, structured answers.

06

Reporting

Dashboards executives actually read. Weekly metric dispatches. A monthly growth memo written in English, not Looker URLs.

GEO · AI-SEARCH

Four surfaces the LLM search layer actually reads

Generative search changed the rules between searcher and link. We prepare the site on four distinct surfaces so the AI engines can resolve the brand, quote the content and return it as a citation.

Entity graph

Identity the search layer can resolve

Name disambiguation, schema.org coverage, knowledge-panel alignment, cross-domain author attribution. A brand is an entity first, a website second.

Citation-worthy

Content a model will quote, not paraphrase

Specific numbers, primary sources, named authors, publication dates, direct answers to direct questions. Reduces hallucination risk and lifts citation rate.

Structured answers

Format that survives summarisation

Definition blocks, comparison tables, step lists, FAQ with semantic markup. The shape LLMs extract cleanly when paraphrasing the page.

Feed presence

Where generative engines index from

Sitemap + hreflang, data feeds, partner syndications, wiki / Wikidata footprint, crawl access from AI user agents. Signals that seed retrieval.

Tracking is a contract

Before dashboards, before attribution, before experimentation — there is an event schema. The table below is the contract we write before any instrumentation lands.

See event schema ↘

EVENT CONTRACT

Eight events every growth build agrees on

The canonical event schema we start every engagement on. Names match Google's recommended ecommerce events for compatibility; parameters are explicit so the warehouse schema stays stable as the product evolves.

Event Parameters Fires on
page_view path · referrer · utm_* Every page
view_item item_id · price · variant PDP
add_to_cart item_id · qty · value · currency Cart actions
begin_checkout value · items · coupon Checkout start
purchase transaction_id · value · tax · shipping Paid
lead lead_id · source · score B2B forms
sign_up method · plan · seat_count SaaS onboarding
feature_used feature_id · workspace_id · role Product

STACK

Tracking · warehouse · SEO · experimentation

Default toolkit across the four lanes growth engineering operates on. Substitutions happen per engagement.

Tracking

  • Google Tag Manager · Server-side GTM
  • RudderStack · Segment · Jitsu
  • Snowplow · event schema registry
  • Posthog · Mixpanel · Amplitude

Warehouse

  • BigQuery · Snowflake · Redshift
  • dbt · Fivetran · Airbyte
  • Reverse ETL (Hightouch, Census)
  • Lightdash · Looker · Superset

SEO · GEO

  • Screaming Frog · Sitebulb
  • Ahrefs · Semrush · Sistrix
  • Schema.org · structured data
  • Log analysis · Lumar · JetOctopus

Experimentation

  • GrowthBook · Eppo · Statsig
  • Optimizely · VWO (legacy)
  • Bayesian + frequentist
  • Cohort & sequential testing

Adjacent disciplines

Where growth connects

Product

SaaS Product Development

The product surface that generates the events. Tracking that outlives a re-platform is wired into the schema from day one.

 Content

Web & Content Platforms

The content layer that the AI-search engines read. Headless CMS, multilingual routing, CWV budgets.

 Revenue

CRM & Revenue Operations

Where the events turn into pipeline, quote and closed revenue — the downstream of every tracking build.

 Reporting

Business Intelligence

Dashboards the executive team reads. Decision-ready reporting instead of Looker URLs in Slack.
Track · attribute · compound

Treat growth like a system, not a series of experiments

Share the current stack, the missing signals and the channel mix. We come back with an event-schema draft, attribution architecture and 90-day experimentation plan inside ten working days.

Start a growth engagement Or go to BI