Sales Funnel Optimisation

sackett.ai developed a digital twin which automatically tracks 4000+ relationships for a small-medium business.

Key benefits:

  • Reduced cost to serve: customers segmented by funnel state for more efficient communications.
  • Increased retention and growth: underserved high-value accounts flagged for immediate attention.
  • Improved customer experience: customers segmented by journey types for tailored experiences.

Our client's business had experienced rapid growth and keeping up with customer communications had become a challenge. CRM datasets were becoming increasingly out of date and incomplete.

Approach:

  • Automatically track sales funnel states (aware, interested, intending, trialling, advocating).
  • Rank accounts by growth rate and total revenue, and flag customers not yet assigned an account manager.
  • Reconstruct thousands of customer journeys for detailed analysis.

The team designed and built a digital twin of the sales funnel which automatically processed email communications, sales orders and invoices, then calculated a sales funnel state for each customer. Customers were then ranked by revenue growth and total revenue, and dormant customers were flagged for follow up. A data quality report identified fields of missing data in the CRM and highlighted underserved high value accounts for attention by the sales team.

Sales Funnel Monitoring

100%Funnel statesConversion ratesTechnology solutionDaily scan for marketing and sales messages sent to 4000+ leadsDaily scan for email correspondence from all 4000+ leadsTopic modelling used to detect intent to purchase in emailsEmails matched with ERP and CRM records using probabilityInformation products (e.g. lists of key accounts, alerts) implemented in Power BIAwareEmail from business to customer observedInterestedEmail from customer to business observedIntent to purchase observedIntendingTrialling1-3 purchasesAdvocatingHigh growth and/or revenue40%30%20%5%

Customers move through the sales funnel from top to bottom. Emails, invoices and other data can be used to observe state transitions and conversion rates.

Key components of the solution were fusing 5+ structured and unstructured datasets and storing resulting data products in a data warehouse; analysing 500,000+ email communications using language models; generating communications summaries with gen AI; and presenting aggregated information products using Power BI. The client required we take a multi-cloud approach including AWS SageMaker, Microsoft Azure and and MLFlow.

Good Data and AI doesn't cost the Earth

Despite the powerful capabilities of this solution, the cost to run was less than $500 per month and development time was less than 3 months.

Would you like some help with sales funnel optimisation? Don't hesitate to send us an email.

Back to case studies