consumption-usage-forecasting

Amit Sasturkar is the co-founder & CTO at BoostUp, where he works on applying AI to the problem of sales forecasting by analyzing a variety of data sources including CRM, call, calendar, email, and third-party data. Amit is a seasoned founder who also co-founded OpsClarity, an intelligent APM for modern web-scale applications. Amit has spent several years at Google Search and Yahoo Search, building large-scale AI systems and backend systems, and has several research papers and patents in web-scale data mining, anomaly detection, and operational data analytics.

Consumption-based pricing models, where customers pay for exactly what they use, are rapidly gaining traction across industries.

Software as a Service (SaaS) and Infrastructure as a Service (IaaS) - where SaaS provides cloud-based applications and IaaS offers virtualized computing resources - both significantly benefit from this pricing approach. Adopting these models is advantageous for businesses and their clients alike, as it aligns costs directly with usage, enhancing transparency and scalability.

For B2B XaaS companies, accurately forecasting customer consumption patterns through consumption-based forecasting is becoming table stakes. Unlike traditional fixed licensing models, consumption forecasting allows businesses to precisely predict future usage and align resources while providing customers with flexible, pay-as-you-go pricing. 

By nailing their consumption-based forecasting, companies can drive sustainable growth, maximize customer lifetime value, and maintain a critical edge in rapidly evolving markets.

The Rise of Consumption-Based Pricing Models

Consumption-based or usage-based pricing models are rapidly gaining traction, especially in the software, digital infrastructure, and services realm. In 2023, 61% of SaaS businesses were already utilizing usage-based pricing models at least in a tiered approach, with full consumption-based models also becoming more popular.1

Unlike traditional fixed licensing models, these flexible models allow customers to only pay for what they use, aligning costs directly with consumption.

For XaaS companies, consumption-based pricing provides a key competitive edge by offering customers more flexibility, scalability, and cost-efficiency. It enables capturing a wider customer base – from small businesses to enterprise clients – by lowering the barrier of large upfront payments. Customers get to trial products easily and scale usage up or down as needed.

Examples

  • SaaS
    • OpenAI: Priced based on number of tokens.
    • Zapier: Payment is calculated based on the number of tasks automated.
  • Cloud Computing and Storage
    • Google Cloud: Pay for what you use.
    • Dropbox: Tiered storage subscriptions.
  • Manufacturing
    • Intel: Chips on Demand. Pay is based on consumption.
    • Stryker: Flexible medical device access. Payment aligns with usage levels.

The shift is also being driven by advancements in cloud computing that make delivering services and tracking usage simpler. Software vendors can tap into new revenue streams by monetizing offerings that were previously difficult to package and sell.

However, transitioning from fixed upfront pricing to variable, usage-based models poses challenges. Revenue streams become less predictable, requiring robust forecasting capabilities. Pricing, packaging, and billing must adapt to complex consumption metrics. Sales compensation and partner incentives may need restructuring.

While overcoming these hurdles demands strategic adjustments, the benefits of consumption-based pricing are making it an increasingly viable go-to-market strategy across industries. By offering the agility today's customers demand, forward-looking companies are well-positioned to drive sustainable growth.

What is Consumption Forecasting?

Consumption forecasting, also known as usage forecasting, is the capability to accurately predict future revenue of pay-as-you-go or consumption-based business models. It involves forecasting how quickly customers will consume resources against a contract or purchase order, rather than projecting a fixed amount.

This is in contrast to traditional opportunity or account-based forecasting, where revenue is derived from set contract values for a product or service over a defined period. With consumption-based forecasting, revenue realization depends on tracking actual customer usage over time.

For example, a provider charging per gigabyte of storage consumed each month would need robust consumption-based forecasting capabilities. Forecasts must account for potential fluctuations in each customer's usage demands, new customer acquisition, and churn rates.

As more businesses shift towards flexible consumption pricing models, mastering this type of forecasting has become critical for sales planning, resource allocation, and maintaining predictable revenue streams. Accurate usage-based forecasting enables companies to align production capacity, budgets, and growth strategies to dynamic customer behaviors.

The Importance of Accurate Consumption Forecasting

Accurate consumption forecasting is absolutely essential for companies operating under flexible consumption-based models. It serves as the basis for sound business planning and strategy for the entire company.

From a revenue perspective, reliable consumption forecasts provide much-needed predictability of expected revenue streams. Without this visibility, companies risk over- or under-investing in critical areas such as product development, sales, marketing, and support. Inaccurate forecasts can lead to missed growth opportunities or cash flow problems.

In addition, by precisely projecting customer usage, companies can proactively adjust their production capacity, whether it's cloud computing resources, inventory levels, staffing needs, or other capital expenditures. This maximizes operational efficiency and cost management.

On the sales side, insight into historical and projected consumption patterns is invaluable for identifying upsell, cross-sell, and customer retention opportunities at the account level. If a customer consistently increases their usage, it presents an opportunity to offer incentives for higher usage tiers, such as price breaks or enhanced service levels. 

 Conversely, unusually low usage can be a warning sign of potential churn risks. With ample lead time, sales teams can initiate customer success programs and incentives to reinvigorate adoption.

However, inaccurate consumption forecasts can severely disrupt these processes. Manufacturers could overproduce unwanted inventory. Cloud service providers could underutilize their resources, resulting in less revenue. Customer relationships inevitably suffer from reactive rather than proactive account management.

For any business operating on the consumption-based model, precisely predicting demand patterns is mission-critical for maintaining profitability, competitiveness, and sustainable growth over the long term.

Challenges of Consumption-based Forecasting

Despite its importance, accurately forecasting consumption poses some significant challenges that companies must overcome.

Data Silos and Unstructured Data

For many businesses, critical consumption data remains scattered across multiple siloed systems like CRMs, ERP platforms, and data warehouses. This data exists in unstructured formats, requiring extensive cleansing, transformation, and integration before it can realistically fuel forecasting models. Without a centralized, standardized data foundation, forecasts are likely to be highly inaccurate.

Evolving Consumption Behaviors

Especially with digital services and subscription models, customer behaviors are not static - they constantly evolve, based on product updates, emerging use cases, and changing business needs. Forecasting models must be able to continuously adapt to these fluid consumption patterns to maintain precision. 

Integrating with tools that monitor product usage and customer satisfaction provides vital insights into these evolving behaviors. Such integrations facilitate more accurate forecasting, particularly regarding subscription renewals. 

Sales and Finance Alignment

Accurate usage forecasting requires a tight collaborative feedback loop between sales teams closest to customer demand signals and financial teams responsible for precise revenue projections. Lack of communication and process misalignment between these critical functions leads to distorted forecasts.

Manual Forecasting Limitations

Historically, consumption forecasting has relied heavily on manual number-crunching by analysts - consolidating data from multiple sources into spreadsheets to develop forecasts through laborious pivot tables and formula-based calculations. This archaic process is highly inefficient, introduces human error, and cannot properly handle the complex multidimensional inputs required for precise forecasts.

External Variable Impacts

Customer consumption patterns rarely exist in a vacuum. A multitude of external factors like economic conditions, seasonal shifts, competitive movements, and more can dramatically influence demand. Effectively accounting for these variables requires sophisticated modeling capabilities that most companies lack.

Strategies for Effective Usage Forecasting

To overcome the challenges of accurate consumption forecasting, companies must adopt a multi-faceted approach spanning technology, processes, and organizational capabilities.

Unified Data Foundation

Creating a centralized, continuously updated data repository integrating all relevant consumption signals from source systems is table-stakes. Data warehousing and ETL tools are critical for cleansing, standardizing, and structuring large volumes of consumption data for analysis.

Advanced Forecasting Algorithms

Traditional forecasting methods relying on linear regression models cannot adequately capture the complex dynamics, like behavior changes, seasonal variations, service or product availability, and economic conditions, influencing customer consumption patterns. AI and machine learning algorithms that can automatically explore diverse variable combinations, detect subtle correlations, and continuously retrain themselves as conditions change are essential.

External Data Integration

Feeding forecasting models a diverse range of external data inputs - economic indicators, market trends, competitive intelligence, etc. - allows them to more precisely factor in environmental variables impacting demand. APIs, data parsing, and orchestration capabilities facilitate this external signal ingestion.

Cross-Functional Collaboration

Bridging long-standing operational silos between sales, finance, product management, and other stakeholder teams is key. Implementing a central forecasting framework with continuous communication feedback loops ensures forecasts incorporate end-to-end organizational insights.

Augmented Human Intelligence

While automation is critical, human expertise remains indispensable for activities like evaluating outlier forecasts, determining drivers behind forecast fluctuations, and calibrating models. Solutions that seamlessly combine machine intelligence with human decision-making produce the most reliable forecasts.

Continuous Monitoring and Recalibration

Customer behaviors and market dynamics are fluid. Organizations must monitor forecast accuracy on an ongoing basis and recalibrate forecasting models based on the latest demand patterns to maintain precision over time.

Organizational Change Management

Technology alone is insufficient. Companies must develop internal capabilities through training programs, documented best practices, and culturally embracing the pivot toward data-driven, AI-assisted forecasting processes.

Implementing a holistic strategy addressing each of these areas provides a solid foundation for mastering the complexities of usage-based forecasting. Those who succeed gain a formidable competitive edge through heightened operational agility and customer responsiveness.

The Future of Consumption-Based Forecasting

The future of consumption-based forecasting is being radically reshaped by advances in artificial intelligence (AI) and machine learning (ML) technologies. These solutions are enabling unprecedented automation and accuracy in predicting dynamic customer consumption patterns.

Traditional rules-based forecasting methods are becoming obsolete in the face of multi-dimensional, ever-changing demand signals. AI/ML models have a distinct advantage - they can ingest vast datasets across multiple variables, automatically identify complex correlations, and continuously re-train themselves as new data arrives. This allows for much higher precision forecasts that quickly adapt to evolving consumption trends.

We're seeing a new breed of AI-powered forecasting solutions emerge, purpose-built for the unique challenges of consumption-based business models. These leverage machine learning to crunch billions of data points - from product usage traces to voice/email sentiment - to reveal granular demand drivers. Augmented with human domain expertise, this intelligent forecasting gives companies a strategic edge through heightened demand visibility.

How BoostUp Enables Consumption-based Planning and Forecasting

  • Unified Data Ingestion, Storage, and Analysis: Gain a holistic view of your XaaS revenue stack as BoostUp seamlessly ingests and aggregates data from CRM, email, calendars, call recordings, marketing, and customer success data. Manage and analyze historical data effectively with our comprehensive storage solutions.
  • Advanced Forecasting Algorithms: Enhance your decision-making with BoostUp's AI-based forecasting models, which are customized for your business. These models integrate diverse signals, including rep behavior, historical conversion rates, and CRM data, ensuring continuous adaptation and accuracy.
  • Cross-Functional Collaboration: Standardize and extend BoostUp dashboards beyond the sales team. Integrate Finance, Marketing, BizOps, and other departments to share the same level of data visibility and integrity, promoting a unified approach to business intelligence.
  • Insight Generation: Use BoostUp to analyze key deal risk factors and sales rep performance to generate actionable insights that you can use to refine your consumption trend strategies.
  • Adoption Tracking: Monitor new customer adoption rates effectively using BoostUp's user-friendly interface, helping you understand engagement and uptake in real time.
  • Process Optimization: Optimize your business processes with BoostUp’s tools for rigorous, consistent reviews and forecasting, driving operational consistency at all levels.
  • Predictability Improvement: Achieve more reliable forecasting with real-time pipeline analytics and quarterly pipeline tracking to improve revenue predictability.
  • Productivity Enhancement: Boost your team's productivity metrics with strategies aimed at increasing win rates, shortening sales cycles, and improving overall sales rep performance.

As these AI/ML-powered forecasting capabilities become mainstream, we'll see accelerated market disruption. Companies embracing intelligent forecasting will gain a significant competitive advantage through maximized revenue realization, resource optimization, and Customer 360 insights fueling stellar client experiences.

References

Kyle Poyar, Sanjiv Kalevar, Curt Townshend - THE STATE OF USAGE-BASED PRICING (OpenView), 04. 17. 2024 (link: https://offers.openviewpartners.com/state-of-ubp-second-edition)