Client name changed to protect confidentiality.
The request
What they asked for
Brackenfold wanted its go-to-market team and its leadership looking at the same picture. The ask came in three parts: a planner where the plan for the quarter lives, funnels that show what is actually happening against that plan, and an insights area where leadership can ask the harder questions about revenue, retention and efficiency.
They also wanted it to live where people already work. That meant Gmail and Calendar connected, so the system could see the conversations and meetings that make up a sales motion, instead of asking reps to log them twice.
The problem
What was broken
The plan was one document. Pipeline was another. Retention was a spreadsheet somebody refreshed before board meetings. Each of those was correct on its own, and together they never agreed. A marketing report would show activity going up while the revenue report showed something flatter, and there was no shared definition that explained the gap.
This is the usual shape of marketing that makes noise and never connects to revenue. The activity is real. It just is not tied to a plan, a funnel stage or a retained dollar, so nobody can say whether it worked. The work was to build one place where the plan, the funnel and the money are the same data.
- The plan and the pipeline were never compared on the same screen.
- Meetings and email threads sat outside any reporting.
- Retention and cohort views were manual and out of date by the time they were read.
- Cost and profitability sat far from the channels that drove them.
What we built
The system, piece by piece
Structure
Three areas, one data set
The product splits into planner, funnels and insights, and all three read from the same underlying data. That is the design decision everything else depends on. A target set in the planner is the same number the funnel measures against, and the same number the insight dashboards use when they talk about efficiency.
Planner
Where the go-to-market plan is written down as targets and owners, so the plan is data and not a slide.
Funnels
Stage-by-stage views of the pipeline, measured against what the planner says should be happening.
Insights
The leadership layer: revenue, retention, efficiency and profitability, built on the same records.
Insights
Eight dashboards, each one a question
Each insights dashboard exists because leadership asks a specific question, and each one answers it without a second spreadsheet.
| Dashboard | The question it answers |
|---|---|
| GTM overview | Are we on plan, and where are we not? |
| Revenue movement | What moved revenue this period: new, expansion, contraction or loss? |
| Revenue retention | Of the revenue we had, how much did we keep? |
| Cohorts | Do customers who joined at different times behave differently? |
| Sales efficiency | What does it cost us in effort and spend to add revenue? |
| Channels | Which channels produce revenue and not only activity? |
| Unit economics | Does a customer pay back what it cost to get them? |
| Profitability | After costs, what is actually left? |
Connections
Gmail and Calendar, so the work is not logged twice
Gmail and Calendar are connected so the system can see the activity that surrounds a deal. Meetings and threads attach to the accounts they belong to. Reps do not retype what they already did, which is the most reliable way to keep a CRM-style record alive.
Connecting a mailbox is a trust decision. The planner and funnels use that activity as context about accounts. They are not a mail client.
MCP server
The same data for AI tools, with the same permissions
An MCP server exposes the dashboards’ data to AI assistants. Someone can ask a question in their own tool of choice and get an answer grounded in the real numbers, rather than a plausible guess.
The important rule is that the MCP server is another door into the same system, not a side channel. It reads the same data, uses the same definitions and respects the same access limits as the dashboards. If a person cannot see a figure in the product, the assistant cannot read it for them.
Trade-offs
Decisions that shaped it
A reporting product like this is easy to over-build. These are the calls that kept it pointed at revenue.
Plan as data
Putting the plan in the planner, instead of in a slide, costs the team a little discipline up front. In return, every funnel and dashboard can say how far from plan it is.
Questions over charts
Each insight dashboard exists to answer a named question. If a chart could not be tied to a decision it was left out.
Retention next to acquisition
Revenue retention and cohorts sit beside the funnel because new revenue that does not stay is a different story from revenue that does. Showing only acquisition flatters the picture.
Costs in the same place as revenue
Unit economics and profitability are part of the product, not an annual finance exercise. Marketing that looks good before cost can look very different after it.
One door for people and assistants
The MCP server reuses the product’s definitions and access rules. That is slower to build than a shortcut and much safer to run.
Handover
Who runs it
The GTM team owns the planner and the weekly funnel review. Leadership owns the insight dashboards and the questions they are built around. Changing a target or a definition is a product setting in the plan, not a conversation about whose spreadsheet to edit.
Operations
How it runs day to day
The planner is where the week starts. Targets and owners are set there. The funnels are where the team checks in during the week: which stage is behind, which accounts have gone quiet, whether the meetings on the calendar match what the pipeline says is coming.
Leadership works from the insights dashboards on a longer rhythm. Because they use the same records as the planner and funnels, a monthly review is a conversation about the numbers instead of an argument about whose spreadsheet is right. The MCP server covers the in-between moments, when someone wants a quick answer without opening a dashboard.
Measurement
What the system measures
These are the areas the dashboards track. They are definitions, not results.
- Plan against pipeline, in the GTM overview
- Revenue movement, including new, expansion, contraction and loss
- Revenue retention
- Customer cohorts over time
- Sales efficiency
- Performance by channel
- Unit economics
- Profitability
Scope boundaries
What we chose not to build
- A forecasting black box
- The numbers are traceable to records. A figure nobody can explain is worse than no figure.
- Activity scoreboards
- Emails sent and calls made are context. The dashboards are organised around revenue, retention and cost.
- An AI path that skips permissions
- The MCP server reads through the same access rules as the product, with no special view for assistants.
- Another place to log activity
- Gmail and Calendar are connected so reps do not type the same thing into a second tool.
Outcome figures
Outcome figures are held back for confidentiality and will be added when the client approves publication.