Process daily revenue reports
Companies typically wait until the end of the month for revenue reports, limiting their ability to make timely decisions. With Metronome’s invoice breakdown export, that’s no longer the case. Companies can now access daily revenue insights, transforming the way they manage their finances. This faster, more detailed reporting dramatically shortens the feedback loop, empowering businesses to make informed decisions on a daily basis.
This page walks through examples of how the Metronome invoice breakdown export powers decisions like resource allocation and upsells.
Invoice breakdown export is currently in beta. It’s available for production use, but you may encounter product limitations. Breaking changes may occur.
How invoice breakdown export works
Metronome’s invoice breakdown export gives a detailed view of how individual line items evolve throughout an invoice period. This export breaks down each invoice line item on a daily basis, allowing you to see the quantity, unit price, total charges, and any credits or commitments applied to each item for a given day. Use this level of granularity to precisely track daily product usage and revenue.
You can schedule this export to transfer to your warehouse once per day. This transfer populates two tables:
invoice_breakdown.invoices, invoices for each of your customers, broken out by day. You receive a row per invoice, per day, for each day of the month in each snapshot. Meaning, if today is Oct 9th, you should expect data from Oct 1–Oct 8 in the most recent snapshot.invoice_breakdown.line_items, line items for each of your invoices, broken out by day. You receive a row per line item, per day, for each day of the month in each snapshot.
Each time the schedule runs, each invoice and line item is recomputed. As a result, you should expect day-over-day totals to change as new usage comes in. Consider the fluidity and dynamic nature of the models when choosing which workflows to power off of them.
Invoice breakdowns don’t enforce the same finalization controls as standard invoices in Metronome. If you send backdated usage after an invoice is finalized, this new usage gets reflected on the corresponding invoice breakdown. This can lead to discrepancies between finalized invoices and invoice breakdowns.
For more information on the invoice breakdown dataset schema, see data export.
To enable invoice breakdown data exports, contact your Metronome representative.
Resource allocation example
This example demonstrates how exports can offer valuable insights into product usage and customer spending.
In this example, your company, NeuraTech, is a cutting-edge AI solutions provider that leverages advanced machine learning algorithms to deliver personalized insights for its clients. Due to the complexity and scale of your AI models, you require vast amounts of compute and storage to process large datasets in real time. This infrastructure is critical to providing seamless, high-performance AI services, allowing businesses to harness the full potential of NeuraTech’s artificial intelligence for their operations.
A key concern for you is effectively managing your hardware resources to optimize for associated infrastructure costs and end user experience. To do this, you need to understand how margins change across different products during times of peak and low usage. For example, if a certain model receives no usage during a period of time but the underlying infrastructure remains constant, the product can fall into the red for that period.
Backed by the invoice breakdown tables, you can generate a time series graph that displays aggregate usage broken down at the product level across daily windows. Effectively, this visualizes top-line revenue for each product. Here’s an example of what this graph could look like:
From the graph, you can derive two key insights:
- Significant increase in revenue for Model A
- Significant decrease in revenue for Model B
Given the difference in revenue, you may want to optimize for the experience of users leveraging Model A as opposed to B. For instance, you may want to reduce the latency of responses returned from Model A. To accomplish this, reallocate resources from Model B to Model A.
This scenario highlights how you can make dynamic resourcing decisions to impact product margins and user experience. However, you can imagine a use case where a pattern emerges. Over time, you might realize that this isn’t a sudden increase in usage for Model A, but actually a pattern. Armed with this insight, you can proactively schedule appropriate resourcing to optimize margins and user experience of the product over time.
GTM upsells example
In the previous example, your NeuraTech’s engineering team leveraged invoice breakdowns. This next example explores how your GTM team can derive value from the same dataset.
Instead of aggregating usage across all invoices, create a dashboard that shows a specific customer’s spend overlayed over their remaining commit balance:
The customer’s usage increased significantly in the current month, burning through the majority of their commit balance. Given that the GTM team doesn’t need to wait until the end of the month, when the customer may go into overage, they can proactively reach out to the customer to discuss their usage pattern and purchasing a new commit.
For more insights, you can segment the customer’s usage further to see if the increase is associated to a specific product. If so, you can offer a discounted rate for the product as an incentive for the customer to purchase the new commit.