How We Designed
Our GTM Tech Stack.
What we're using and why — chosen for our objectives, modularity, cost-efficiency, and team adoption. Your stack will vary, but the principles stay the same.
Layer by Layer
Each tool maps to a function in our three-layer GTM system. Monthly costs shown are typical starting points.
Why These Choices
Every tool earns its place. Here's the reasoning behind each layer of the stack.
Start at the End
Before looking at tools, we looked for clarity on what we wanted to accomplish. What value do we offer and to whom? How will we visualize the future and make decisions? Who is the team and how would their time be used most effectively? With a modular architecture, we could plan for our current reality, and grow into our future.
We then prioritized our initial engagement activities. Some were critical and specific, deserving more investment, others could be optimized or added later. Our interaction needs would change, guaranteed. Our top priorities:
Single source of truth over tool sprawl
PostHog was an easy choice to provide a single source of truth, consistent schema and tracking and visualize insights into the complete revenue generation picture. It provides the warehouse, CDP, analytics and visualization we needed to be the foundation of a stable data core. Our team has no dedicated sales force, so a CRM like Hubspot or Salesforce isn't necessary, and PostHog provides a foundation for one when needed. We would start at zero cost and scale to thousands of customers and millions of visitors before we exceeded $1000/month.
For AI in engagement and content creation, scaling and accessing our existing knowledge base at an affordable cost was crucial. Pinecone vector database underpinning AI as a force multiplier gave us the guardrails to control costs and keep AI focused on our objectives and our customers.
Composable logic that tolerates change
Our workflow needs were simple -- a few stable inbound workflows, a content gen workflow, and a social publishing workflow. It came down to Zapier Canvas and Make for our budget, and the decider was time. Zapier Canvas allowed us to spend less time creating and updating workflows, provided the connectors and AI features we needed for interpreting signals, making it worth the step up in price from Make.
A stable data core means we can change the orchestration layer when needed. Costs will increase, needs will change, and we will eventually move to a more specialized tool like Default that can handle more complexity with even less time commitment. At that time, the decision will be easy.
Customer touchpoints with built-in feedback loops
Our interaction needs will change more rapidly and frequently than anything else in our stack. Modularity is not an option here, and priorization, ruthless. Knowing our current needs guided us to email and chat as the top priorities.
With our data core feeding events and data, and our orchestration layer integrating tools, leveraging AI and handling most of the decision making, email tool needs got more focused. We needed high deliverability, simple automation and custom events at an affordable cost, and Loops fit the bill. Other tools with higher price tags have more features, but we don't need them now, and won't pay for them until we do.
AI chat was the largest bet we took with our time. Getting it right for our visitors had big upside in goodwill, conversion, and as a showcase of the convergence of all layers of our stack. A custom build was the best way to achieve those goals.
[Build your own stack]
Every team is different. Let us design the right architecture for yours.
Have questions about building the right stack for your team? Reach out and we'll talk through it.