Infrastrutture AI Infrastrutture AI
AI Callers and RAG Chatbots Client model: MangoTeq (UK)
Version 1.0 · August 2026

MangoTeq — Acquisition KPI Model

A working model of the AI driven acquisition funnel: paid traffic on Meta and Google, a WhatsApp chatbot that captures the opt in, an AI caller and chatbot that book the consultation, and the consultation that closes the client. Move any assumption on the left and every figure on this page recalculates.

Example scenario

Illustrative figures, not actuals. Every number on this page is produced by the assumptions panel, not by measured MangoTeq performance. The model exists to agree on which metrics we track and how they relate, before any spend goes live.

Monthly performance

One steady month at the assumptions on the left. Stage counts are whole units, so every figure below can be checked against the one above it.

Revenue booked per month

£40,500

9 new clients at £4,500 average deal value

ROAS

13.5x
revenue per £1 of ad spend

Cost per acquisition

£333
7.4% of deal value

Ad spend

£3,000
Meta and Google, per month

Leads

250
at £12.00 per lead

Cost per lead

£12.00
blended across both channels

Opt ins

138
55.2% of leads

Meetings booked

48
34.8% of opt ins

Show rate

70.8%
34 of 48 booked meetings attended

New clients

9
26.5% close rate on consultations

Cost per meeting

£62.50
per meeting booked

Cost per acquisition

£333
ad spend per new client

ROAS

13.5x
£40,500 revenue on £3,000 spend

Acquisition funnel

Bar length is the stage volume. The figure under each bar is the conversion into the stage below it.

Paid traffic on Meta and Google
Campaign spend is the funnel input, not a stage volume, so it is not drawn to the same scale. £3,000
Show data table

Six month projection

Steady state: the same monthly spend and the same conversion rates repeat every month.

Cumulative revenue against cumulative ad spend

Both series are in pounds on one axis. The shaded band between them is the cumulative gross contribution, that is revenue booked minus media cost, before delivery and platform costs.

  • Cumulative revenue
  • Cumulative ad spend
  • Gross contribution
Show data table

Cost per funnel stage

The same monthly ad spend divided by the volume that survives to each stage.

What one unit costs at each stage

Cost compounds down the funnel: the further a unit travels, the more media spend sits behind it. Stage colours match the funnel above.

Show data table

How the model calculates

Seven levers, everything else derived. No hidden constants.

Volume chain

Leads = ad spend / cost per lead, then each stage multiplies the stage above it by its rate and rounds to whole units. Every stage figure on the page is derived from the rounded figure above it, so the funnel always ties out.

Cost and return

Cost per stage = ad spend / stage volume. Revenue = new clients × average deal value. CPA = ad spend / new clients. ROAS = revenue / ad spend. Revenue is booked value, not collected cash, and excludes delivery cost, platform fees and VAT.

What the AI layer moves

The chatbot moves the opt in rate, the AI caller and the chatbot together move the booking rate and the show rate through reminders and reschedules. Cost per lead is a media variable, close rate stays with the MangoTeq sales team.

What to replace with real data

After four to six weeks live, cost per lead, opt in rate and booking rate come from platform and CRM data. Show rate and close rate need one full sales cycle before they are reliable. Until then, treat every figure here as a planning assumption.