The real stack, not a wish list
The marketing tools I use to turn clicks into revenue you can trust
The exact stack I run as a Fractional CMO (Fractional Head of Growth), what each tool earns its slot for, and the one number I trust when the tools disagree.
Operated growth at
- Elementor
- Riverside
- Intel
- Elementor100x$200K to $20M ARR as acquisition lead
- Riverside+337%MRR growth driven as a growth operator
- Across engagements$100M+ad budgets managed across paid social and search
In short
The marketing tools I use are not a wish list. They are the exact stack I run as a Fractional CMO (Fractional Head of Growth), every week, on real budgets. Every tool here earns its slot by connecting a click to a dollar, or it gets cut. That is the whole filter.
I do not pick software because a vendor sponsored a newsletter. I pick it because it answers a revenue question I cannot answer any other way. If a tool cannot tell me where the money came from or where it leaked, it does not survive my stack.
On affiliate links: a few tools below are marked Partner link, and tools I built myself are marked In-house. I only ever link tools I actually run, and a commission never changes what I recommend. Full disclosure.
The stack I actually use
Every tool I run, grouped by the job it does
These are the tools I use daily to build growth systems for clients. If something isn't here, I either tried it and dropped it, or haven't needed it.
Last updated: May 2026
AI & Automation
Claude Code
AI coding agent for building automation pipelines, SEO audits, content production, and data analysis.
How I use it: runs my entire SEO audit engine, content pipeline, and client reporting automation.
n8n
Open-source workflow automation. Self-hosted. 62+ workflows running client operations 24/7.
How I use it: marketing ops, attribution pipelines, alert systems, CRM sync, content distribution.
Firecrawl
Web scraping and content extraction API. Handles JS-rendered pages.
How I use it: competitor analysis, SERP scraping, content auditing at scale.
ChatGPT / Codex
OpenAI's assistant plus Codex for agentic coding. My second model alongside Claude.
How I use it: research, fast drafts, and code generation when I want a second opinion.
Grok
xAI's model, wired to live X and the open web for current, fast research.
How I use it: real-time trend checks, social listening, and quick fact pulls.
Opencode
Open-source AI coding agent that runs in the terminal against any model.
How I use it: self-hosted scripting and automation where I want full control of the model.
Browserless
Headless Chrome as a service. Renders JavaScript pages that plain scrapers choke on.
How I use it: pairs with Firecrawl for JS-heavy SERP and competitor scraping.
Hermes
My self-hosted agent layer on the VPS. Routes each job to the right model and tool.
How I use it: the brain behind scheduled research, audits, and content runs.
Sticklight
Vibe-coding environment for building apps and internal tools by prompting instead of boilerplate.
How I use it: fast prototypes of client dashboards, landing-page tests, and internal marketing tools.
SEO & GEO
Google Search Console
The only SEO tool that shows real data. Everything else is an estimate.
How I use it: weekly keyword tracking, click-through analysis, indexation monitoring.
Rank Math
WordPress SEO plugin. Schema markup, meta tags, sitemaps, breadcrumbs.
How I use it: on-page SEO management, JSON-LD schema, llms.txt generation.
Screaming Frog
Site crawler for technical SEO audits. Finds broken links, duplicate content, missing tags.
How I use it: pre-engagement technical audits, ongoing crawl monitoring.
Bing Webmaster Tools
Microsoft's search console. Free index and query data most teams ignore.
How I use it: daily sitemap submission, index monitoring, and Copilot-facing visibility.
Semrush
Keyword, backlink, and competitor research at scale.
How I use it: keyword mapping, content-gap analysis, and tracking competitor moves.
LLM Citation Tracker
My own llm_citation.py engine. Checks if AI engines cite a page, and for which prompts.
How I use it: weekly GEO tracking across ChatGPT, Perplexity, Gemini, and Google AI answers.
Analytics & Attribution
Google Analytics 4
Web analytics. The industry standard, warts and all.
How I use it: traffic analysis, conversion tracking, audience insights.
Mixpanel
Product analytics. Event-based tracking, funnels, cohorts, retention.
How I use it: SaaS funnel analysis, feature adoption, revenue attribution.
PostgreSQL
Open-source database. I build custom attribution pipelines connecting ad spend to revenue.
How I use it: marketing data warehouse, custom dashboards, cross-platform attribution.
Google Tag Manager
Container for tags, triggers, and conversion events without shipping code each time.
How I use it: deploying and versioning tracking cleanly across client sites.
PostHog
Open-source product analytics. Events, funnels, session replay, self-hostable.
How I use it: self-hosted funnel and activation analysis when data must stay in-house.
CRM & Lifecycle
Twenty CRM
Open-source CRM. The system of record where pipeline and closed revenue land.
How I use it: the CRM wire in the attribution loop, joining deals back to the click that started them.
HubSpot
CRM, marketing, and lifecycle in one. The default when a client already runs it.
How I use it: pipeline, lead scoring, and lifecycle email for B2B clients on HubSpot.
Klaviyo
Email and SMS built for ecommerce, tied to store and revenue data.
How I use it: abandoned-cart, post-purchase, and winback flows for D2C clients.
MailPoet
Email and newsletters native to WordPress. Light and fast to ship.
How I use it: lifecycle and newsletter sends for WordPress-based clients.
Paid Acquisition
Google Ads
$100M+ managed across Search, Shopping, Display, YouTube, Performance Max.
How I use it: primary paid channel for B2B SaaS and e-commerce clients.
Meta Ads
Facebook and Instagram advertising. Prospecting, retargeting, lookalikes.
How I use it: D2C acquisition, B2C scaling, creative testing at volume.
LinkedIn Ads
B2B advertising. ABM targeting by title, company, industry.
How I use it: enterprise demand gen, ABM campaigns, thought leadership promotion.
CMS & Web
WordPress
The CMS most clients already run. Open, extensible, API-driven.
How I use it: content publishing through its API so changes are repeatable and reversible.
Elementor
Visual website builder for WordPress. I led acquisition there from $200K to $20M ARR.
How I use it: fast landing pages and site builds without waiting on a dev queue.
Infrastructure
Coolify
Self-hosted PaaS. Runs all my services on a single VPS.
How I use it: hosting client tools, automation services, databases, and AI agents.
Cloudflare
CDN, DNS, SSL, security headers. Every client site goes through Cloudflare.
How I use it: performance, security, caching, bot management.
Redis
In-memory data store. Caching, session management, real-time analytics.
How I use it: caching API responses, rate limiting, real-time dashboards.
Hostinger
Affordable managed hosting and VPS. A solid default for client sites.
How I use it: hosting WordPress sites and lightweight client infrastructure.
SearXNG
Self-hosted, open-source metasearch. Private, scriptable, no tracking.
How I use it: programmatic search inside automations without rate-limited APIs.
Want to know what stack I'd recommend for your business? Let's talk.
The problem
Why do your tools report different conversion numbers?
GA4, advertising platforms and CRM can report different objects and attributed credit. Define the event, identity, period and attribution settings before comparing them. A CRM closed-won deal is not automatically collected cash; reconcile it with billing or payment records for that question.
GA4 counts sessions
Review the report scope and selected attribution model. GA4 can measure configured post-signup events; the data available depends on the actual implementation and permissions.
Ad platforms self-credit
Platforms can assign overlapping credit under different windows and definitions. Reconcile the records before aggregating; adding attributed totals can double-count the same outcome.
Your CRM records deals
A recorded deal stage or closed-won value follows your CRM definitions. Check billing or payment evidence for collected revenue, and join only the eligible identities and events that the available data supports.
Which number to trust
The same conversion, counted three ways
Read every number for what it is good at. Trust the warehouse view for the one question that pays the bills: cost per paying customer by channel.
| What you compare | GA4 | Ad platforms | Warehouse |
|---|---|---|---|
| What it counts | Sessions and on-site events | Self-credited clicks and views | Paying customers in billing |
| Counting method | Last non-direct click | Each channel claims credit | Spend joined to revenue |
| Best for | Traffic and on-site funnels | In-platform optimization | Cost per paying customer |
| Trust for revenue | Directional | Inflated | Source of truth |
The stack, by layer
Six layers, one job: connect a click to a dollar
Measurement
GA4 plus product analytics for funnel and activation
Attribution
warehouse view joining ad spend to product and billing data
SEO and GEO
Screaming Frog, Search Console, Bing, AI-citation tracking
Automation
n8n nervous system for spend pulls, reports and alerts
Paid media
direct in Meta and Google Ads, no layer in between
Content and lifecycle
markdown drafts, version control, email automation
The filter for every tool here
Each tool earns its slot by connecting a click to a paying customer. Measurement and attribution come first; everything else is execution on top of a number you can trust.
From click to dollar
The marketing tools i use to go from traffic to revenue
Six layers, in the order a click travels them. Tap any layer to open how I run it, the tools that earn their slot, and the revenue question it answers.
Measurement comes first
Start with measurement, because everything else is guessing without it. My base layer is GA4 for traffic and event tracking, paired with a product analytics platform for funnel work. GA4 tells me what channels drive sessions and conversions. Product analytics tells me what users actually do after they land: signup, activation, the drop-off step that quietly kills the funnel. I instrument events deliberately, name them so they read like English, and I check the data against the source of truth before I trust a single chart. When I took Elementor to 100x ARR, the difference between a good month and a flat one usually showed up in one funnel step, not the headline number. The marketing tools i use for analytics exist to surface that one step fast.
- GA4
- Mixpanel
Attribution is where most stacks break
Attribution is the second pillar, and it is where most stacks fall apart. GA4 gives you a model, but it does not stitch ad spend to revenue cleanly across signup and subscription. So I build a warehouse view: ad platform spend joined to product events joined to billing data. That means raw pulls from Meta Ads and Google Ads APIs landing in Postgres, then a query that answers the only question a CFO cares about: cost per paying customer by channel, by creative, by week. The marketing tools i use here are not glamorous. They are an API client, a database, and a scheduled job. Boring infrastructure beats a pretty dashboard that lies.
Boring infrastructure beats a pretty dashboard that lies.
- PostgreSQL
- Ads APIs
- Twenty CRM
SEO and GEO share one foundation
For SEO and GEO, I run a layered setup. Screaming Frog for the technical crawl: broken links, canonical mismatches, thin pages, schema gaps. Google Search Console and Bing Webmaster Tools for the demand side: what queries actually surface my pages and where impressions hide. Then I add structured data with JSON-LD so pages qualify for rich results and so AI engines can cite them. Search is splitting into classic blue links and generative answers, and both reward the same thing: clean technical health and content that answers a real question. You can read Google's own guidance in the Google Search Central SEO starter guide, which still maps closely to what moves rankings in practice.
- Screaming Frog
- Search Console
- Rank Math
- Bing
- Semrush
Automation buys back the hours
Automation is where I buy back hours. n8n is my nervous system: it connects ad APIs, the database, Slack, and reporting into workflows that run on a schedule without me touching them. A daily pull lands fresh spend and revenue. A weekly job builds the channel report. An alert fires when a cron breaks or a metric drifts. The marketing tools i use for automation are not about looking busy. They are about making sure the data is fresh on Monday morning so the meeting is about decisions, not about why the numbers are stale. Every recurring task that does not need my judgment becomes a workflow.
- n8n
- Claude Code
Paid media: automation reads, a human decides
On paid media, I work directly inside the ad platforms rather than through a layer that hides the levers. Meta Ads Manager and Google Ads, pulled through their APIs for analysis, pushed to manually for the decisions that matter: budget shifts, creative kills, audience tests. I managed $100M+ in budgets by being ruthless about the read-write split. Read everything through the API into one place. Write changes by hand, deliberately, with a clear hypothesis. The marketing tools i use for paid media give me speed on the analysis and discipline on the spend. Automation reads. A human decides.
- Google Ads
- Meta Ads
- LinkedIn Ads
Content and lifecycle close the loop
Content and lifecycle round out the stack. I draft in plain markdown, version it in git, and ship to WordPress or the relevant CMS through its API so changes are repeatable and reversible. For lifecycle email I map the full journey first, then build sequences in the customer engagement platform the client already uses, so every send drives a single action: open, click, do the thing. I drove Riverside +337% Net MRR partly by treating email as a revenue channel with its own funnel, not a newsletter. The marketing tools i use for content and lifecycle are chosen for one trait above all: do they let me trace a message to a result.
I only ever link tools I actually run; a commission never changes what I recommend.
The pattern across all of it is the same. I start from the revenue question, I pick the lightest tool that answers it, and I wire the boring infrastructure so the answer arrives without me asking twice. New software has to clear that bar before it joins the marketing tools i use. If a shiny platform cannot show me cost per outcome, faster than what I run now, it stays out. The stack is a means, not the point. The point is moving a business from traffic to revenue, and these are the tools that let me prove I did.
Work with me
Have me run this stack on your numbers
Diagnostic
$6-8Ktypical range; scoped per engagement
Audit of your stack, data, and attribution gaps
Operator
Pricing and terms agreed for the engagementtypical range; scoped per engagement
I run the stack as your fractional head of growth
AI infra build
from Pricing and terms agreed for the engagementtypical range; scoped per engagement
Custom attribution warehouse and automation
What you walk away with
One number you can trust: cost per paying customer by channel, by week
A warehouse view joining ad spend to product and billing data
n8n automations so the data is fresh and current every Monday morning
A documented stack your team keeps running after the engagement ends
FAQ
Frequently asked questions
What marketing tools i use most often day to day?
GA4 and a product analytics platform for funnels, the Meta and Google Ads APIs pulled into Postgres for attribution, n8n for automation, and Screaming Frog plus Search Console for SEO. The daily drivers are whatever connects spend to revenue. I open the attribution warehouse first every morning, because that is where budget decisions actually get made.
How do you choose which tool to add to the stack?
One filter: can it answer a revenue question I cannot answer now, faster than my current setup? If a tool cannot tie a click to a paying customer or expose where the funnel leaks, it does not get added. I pick the lightest option that works, wire it into a scheduled job, and cut anything that turns into maintenance without a clear payoff.
Before choosing a platform, write a short acceptance test using your own workflow: the input it must accept, the output you need and how you will check correctness. Compare the permission it needs, data you would expose, export options, integration effort and ongoing maintenance as well as the subscription fee. Test the smallest useful scope before making it a dependency. Name an internal owner and an exit path if the tool stops earning its place. A model or vendor leaderboard is not a substitute for this test, especially when the job includes actions that write into your systems.
For a CMS decision, test the pages it actually produces: can a visitor and crawler reach the substantive content, do pages have stable URLs and accurate canonical and language signals, and can editors maintain useful internal links and structured information without guessing? Include redirects, preview and approval needs, and an export path in the requirements. A platform name does not answer these questions; inspect a representative page and the publishing workflow before choosing. Search readiness supports discovery, but no CMS can promise a ranking or an AI citation.
Do you build custom tooling or rely only on off-the-shelf platforms?
Both, deliberately. Off-the-shelf for measurement and ad management, custom for attribution. The hard problem is joining ad spend to product events to billing, and no single platform does that cleanly for every client. So I build a warehouse view with API pulls and Postgres queries. Boring infrastructure that answers cost per paying customer beats a polished dashboard that hides the join.
Where does automation fit among the marketing tools i use?
n8n runs the recurring work: daily spend pulls, weekly channel reports, alerts when a metric drifts or a cron breaks. The rule is simple. Automation reads and refreshes data so it is current on Monday. A human still makes the write decisions: budget shifts, creative kills, audience tests. I never let a workflow change ad spend on its own.
How do these tools support SEO and AI search at the same time?
I run Screaming Frog for technical health, Search Console and Bing Webmaster Tools for demand, and JSON-LD structured data so pages qualify for rich results and get cited by AI engines. Classic search and generative answers reward the same foundation: clean crawlability and content that answers a real question. The stack covers both because they are not separate problems, they are one.
Let's turn this into measurable revenue
Book a 15-min call. I will tell you whether this is your next move, or whether your money is better spent elsewhere.