Our take on MCP
In general, we believe that most apps in the future will be chat-based. Asking for data using natural language is significantly faster and easier for the average marketer. This is particularly true when it comes to analytics where being able to request data and insights using a natural language chat interface is significantly better experience for most marketers than navigating the complex UI of software such as Google Analytics. For more detail see https://growthmethod.mintlify.app/mcp/overview.Ideal customers
Teams that use our MCP server tend to already:- Spend a lot of time working in AI tools like ChatGPT and Claude. MCP enables them to get the benefits of Growth Method for campaign and experiment management but without having to context switch into our app UI.
- Have a relatively modern marketing stack and be comfortable connecting their data-sources and tooling via MCP.
Example workflow
Here is an example of a workflow with the Growth Method MCP server and GA4 MCP server connected to ChatGPT:- “Get details of the pricing page campaign from Growth Method”
- “Use PostHog to compare traffic and conversion data for the /pricing page before and after this campaign went live”
- “Add this traffic and conversion data to Growth Method for campaign analysis”
MCP server best practises
We use industry best practises during our MCP implementation, including:- Our MCP tools feel like commands, they are all clear verbs (verb_object style) with normalised naming conventions to enhance clarity for human and agent
- A small and semantic MCP tool list small semantic and let the schema handle the details. MCP servers implementation that have too many hyper-specific tools makes it harder for models to choose correctly
- Tools always have a single canonical choice, there are no synonyms or duplicates in our implementation
MCP Tools
If you present an LLM with too many tools, or overly complex tools, it may struggle to choose the right one or to use it correctly. As a result, MCP server designers are encouraged to present greatly simplified APIs as compared to the more traditional API they might expose to developers. Cloudflare Engineering
1 Campaign tools
| Tool | Description |
|---|---|
| create_campaign | Create a new campaign in the Idea (Unscored) stage. |
| list_campaigns | List or filter campaigns by stage, owner, date, department, status, etc. |
| get_campaign | Retrieve full campaign details, including stage, hypothesis, build data, results, and analysis. |
| update_campaign_details | Update any editable fields (build, results, analysis). The server enforces stage-based rules internally. |
| change_campaign_stage | Move a campaign between stages: Idea (Unscored/Scored) → Build → Live → Analysing → Complete. |
| get_campaign_results | Retrieve a summary of campaign performance, results metrics, and key outputs. |
2 Comment tools
| Tool | Description |
|---|---|
| add_campaign_comment | Add a comment or update to a campaign. |
| list_campaign_comments | Retrieve comments or discussion history for a campaign |
3 Search tools
| Tool | Description |
|---|---|
| searchDocs | Search all Growth Method documentation, internal guides and help pages. |
| searchTools | Search and retrieve relevant tool definitions so only necessary tools load for the current task. |
MCP technical details
The Growth Method MCP server supports:- Streamable HTTP
- OAuth 2.1 with dynamic client registration at https://mcp.growthmethod.com/mcp
- The meta field as per https://modelcontextprotocol.io/specification/2025-06-18/basic#meta