Backstory's AI Triples Customer Tiering Speed, Slashing Quarterly Task to 3 Days

Automating the Unwieldy: Backstory's Customer Tiering Overhaul

Customer tiering, the process of segmenting a customer base into distinct groups based on value, potential, or strategic importance, is a critical but often cumbersome task for go-to-market teams. Traditionally, this involves extensive data gathering, analysis, and iteration across sales, customer success, and marketing. Backstory, a company focused on customer data infrastructure, recently tackled this challenge head-on, transforming a laborious quarterly exercise into a remarkably swift, AI-powered operation. The result? An entire customer base tiered in just three days, a feat that previously consumed five teams for an entire quarter.

Haya Kamola, Head of Customer Success at Backstory, presented this transformation at SaaStr AI Day, detailing the methodology that dramatically accelerated a process many in leadership roles find deeply familiar. The traditional approach to account tiering is often a manual slog. It involves pulling data from CRM, billing systems, product usage logs, and potentially third-party enrichment tools. This data then needs to be cleaned, standardized, and analyzed to identify key characteristics that define each tier. The process typically requires cross-functional collaboration, with different teams owning specific data sets or analysis segments. This collaboration, while essential for accuracy, introduces communication overhead, review cycles, and potential delays.

For Backstory, this annual or quarterly tiering exercise was a significant undertaking. Five distinct teams were involved, dedicating substantial time and resources. The sheer volume of data, the complexity of defining meaningful segmentation criteria, and the iterative nature of refining these tiers meant the process could stretch for months. This extended timeline meant that by the time the analysis was complete, the market or customer landscape might have already shifted, rendering parts of the analysis less relevant.

The core of Backstory's breakthrough lies in its sophisticated use of its own technology, augmented by AI and robust data connectors. Instead of manual data pulls and spreadsheet wrangling, Kamola's team leveraged custom connectors to pull data from various sources seamlessly. This included not just standard CRM data but also richer signals that provide a more nuanced view of customer engagement and value. Think of it less like a static spreadsheet and more like a dynamic, constantly updated dossier on each customer, compiled automatically.

The AI-Powered Engine: Connectors, Signals, and Iteration

The process began with defining the key performance indicators (KPIs) and signals that would inform the tiering. These weren't just basic metrics like revenue or contract value. Backstory focused on a broader set of indicators that reflect true customer health and potential. These signals were then fed into an AI model designed to identify patterns and cluster customers based on these multidimensional attributes. This approach moves beyond simple rule-based segmentation to a more data-driven, emergent understanding of customer value.

Custom connectors were crucial. They allowed Backstory to ingest data from disparate systems—CRM, product analytics, support tickets, and even external market data—into a unified platform. This consolidated view is the bedrock of any effective segmentation. Without it, teams spend an inordinate amount of time just reconciling data. By automating this data aggregation, Backstory freed up analytical capacity.

The AI component played a pivotal role in the speed and depth of the analysis. Instead of analysts manually sifting through data to find correlations, the AI algorithm could identify complex relationships and patterns that might be invisible to human observation. This allowed for a more sophisticated definition of customer tiers, moving beyond superficial metrics to deeper indicators of engagement, product adoption, and strategic alignment. The AI essentially acted as a super-powered analyst, capable of processing vast amounts of data and identifying subtle clusters.

Kamola highlighted a four-round iteration process, which is critical even with AI. The initial AI output provides a strong baseline, but human oversight and refinement are still necessary. These rounds involved reviewing the AI-generated tiers, validating them against qualitative insights from sales and customer success teams, and making adjustments. The key difference here is that the iterations were on the *output* of the AI, not on the painstaking process of data gathering and initial analysis. This shifted the focus from operational drudgery to strategic validation, a much higher-value activity. Each round of iteration was significantly faster because the underlying data and initial clustering were already robust and readily available.