Skip to main content
We believe that agents and MCP-backed services will become the primary work interface and the front-door to the entire marketing stack over the next couple of years - that the future of user interation will not be in the web browser.
“The work on MCP has completely revolutionized the agentic AI landscape” - Jensen Huang, President and CEO of NVIDIA https://x.com/tadasayy/status/1990817608741106077
Traditional software applications will become predominantly headless, backend platforms that provide data and functions to AI agents via standards such as MCP. The front-end user interface will be reserved for occasional use by power users rather than for day-to-day work. For the average marketer this post-browser world reduces friction and eliminates context switching enabling significant gains in marketer productivity.

Benefits of MCP

With MCP your conversational AI agent “just knows” how to navigate your marketing stack. It can understand your vendor API implementations, chain requests, pass parameters, handle errors and more. There is no need for marketers to understand API specs or which end point to use for a particular task, just describe the task and let your agent handle the rest. At Growth Method we are bullish on MCP as the glue to the marketing stack as it provides:
  • An industry standard
  • First-party vendor maintained servers
  • Standards-based authentication via Oauth
  • A natural langauge interface
  • One-click installation via MCP directories

MCP & marketing campaigns

Here are some examples of using MCP to understand and analyse your marketing campaigns and experiments.
  1. What experiments or campaigns launched have we launched in the last 6 months? [GM]
  2. Which campaigns that were launched in the last 6 months could have affected our site wide conversion rate? [GM]
  3. Give me a breakdown of our marketing campaigns last quarter by channel [GM]
  4. Create a PDF report showing marketing campaigns launched in the last 6 months [GM]
  5. Who on the team shipped the most campaigns last month? Create me a leaderboard
  6. Which page has receives significant traffic but has seen the biggest drop in conversion rate in the last 6 months? Use the Growth Method hypothesis builder to create a new marketing campaign to address this
  7. Are we tracking the right events for the [name] campaign or are there metrics we’re missing?
  8. Get experiment results and data from Eppo for the [name] campaign, write up the results and add them to the Growth Method campaign results
  9. Use the PostHog MCP to get all qualitative feedback from our pricing page and turn this into a hypothesis for a new campaign to run on the page

The analytics use case

For data and analytics work in particular, natural language interfaces provide a significantly better user experience for marketers over traditional user interfaces such as Google Analytics. This makes natural language and MCP the ideal way for marketers to retreieve data for campaign analysis. In this scenario, Growth Method can act as both a server and as a client:
  1. Growth Method MCP Server - Add relevant data to your campaign analysis in Growth Method directly from ChatGPT
  2. Growth Method MCP Client - Add relevant data to your campaign analysis without ever leaving Growth Method

MCP Server vs MCP Client

Consider the following when deciding how to best use Growth Method within your organisation and team.
Growth Method MCP ServerGrowth Method MCP Client
WorkflowUse ChatGPT to get the relevant data from GA4 then use the Growth Method MCP server to add it to your campaign analysisGrowth Method requests the data from GA4 directly and then adds this data to your campaign analysis
Primary user interfaceLLM Client (e.g. ChatGPT)Growth Method
Example user prompt”Get traffic and conversion data for the /pricing page from GA4 then add this information to the pricing page experiment in Growth Method""Get traffic and conversion data for the /pricing page”
Tool auth & managementLLM Client (e.g. ChatGPT)Growth Method
GM RoleBackend, predominantly headlessFront-end and UI
PriorityPrimarySecondary
Note: New MCP protocol versions are released approximately every 3 months as per https://modelcontextprotocol.io/specification/versioning hence functionality is subject to change.