6 months
to deliver Phase 1
1 week → hours
for complex analysis cycles
200–300 MB
Excel workbooks replaced as the main workflow
Phase 2
SaaS transformation underway
The challenge
Critical media decisions were trapped inside scattered tools and fragile spreadsheets
Top of Mind worked with multiple audience, viewership, survey and expenditure sources. Analysts manually extracted data, rebuilt calculations and navigated huge Excel files whenever campaign assumptions changed.
Fragmented sources
Media and audience data arrived in different structures and formats.
Massive workbooks
200–300 MB Excel files created slow and difficult analytical workflows.
Repeated analysis
Any data or parameter change forced analysts to redo part of the process.
Presentation friction
Live questions required searching through multiple files and saved versions.
Our approach
Build a trusted analytical foundation before adding AI
01. Map the analysis lifecycle
Understand the path from raw files to calculations, strategy and presentation.
02. Centralise clean data
Use ETL to standardise sources and store reusable analytical data.
03. Design for analysts
Expose query power visually instead of forcing manual SQL.
04. Preserve explainability
Keep calculations, filters and recommendations visible and defensible.
05. Make exports immediate
Move from analysis to Excel and presentation output without rebuilding work.
06. Add controlled AI
Ground natural-language querying in a known schema and query system.
The solution
One platform for ingestion, analysis, export and AI-assisted strategy
Automated ETL
Clean, standardise and load multiple media-data formats into a central analytical layer.
ClickHouse Analytics
Fast OLAP querying for large-scale filtering, aggregation and comparison.
Visual Query Builder
Drag-and-drop fields, filters, computed columns and aggregates with live preview.
Media Calculations
Reusable TVR, SOM, SOV, SOE, exposure, cost and placement analysis.
Excel & PPTX Output
Export tables and charts directly into client-reporting and presentation workflows.
AI Data Agent
Natural-language questions, schema mapping, query generation and explainable insights.
Technology stack
Decision support
From “what happened?” to “what should we change?”
The system can explain results and suggest where placements may be moved to improve viewership while controlling cost.
Phase 2
Turning internal analytical expertise into a SaaS product
The next phase introduces multi-tenancy, subscriptions, whiteboard analysis, data-scale research and local LLM deployment.
Results and impact
A week-long analytical cycle became an hours-long workflow
Complex analysis cycles dropped from roughly a week to only hours.
Saved queries can be rerun with new months, brands, products and campaign parameters.
Analysts rely less on multiple 200–300 MB Excel workbooks.
Tables, charts and insights move more directly into client presentations.
Successful delivery led directly to a second phase focused on SaaS commercialisation.
Successful delivery of the internal analytics platform led Top of Mind to begin a second phase focused on turning the system into a scalable SaaS product for the wider media-agency market.
Factual project statement — replace with an approved client quotation when available.
Have a similar data problem?