Cost forecasting
Cost forecasting is predicting what cloud spend will be in a future period, from past cost, planned changes and business drivers such as expected orders or customers. Forecasts set budgets, size commitments and warn when spend is heading off plan, and their accuracy is measured as forecast variance: how far actual cost lands from the prediction.
In plain English
Think of a weather forecast. It can't tell you exactly what tomorrow brings, but it's right often enough to decide whether to pack an umbrella, and it gets better when it knows a storm front is coming.
Formula and worked example
Forecast variance = (actual cost − forecast) ÷ forecast
- Forecast
- = the cost predicted for the period, before it started
- Actual cost
- = what the period cost once billed
| Forecast for March | $110,000 |
|---|---|
| Actual March cost | $115,000 |
| Difference | +$5,000 |
| Forecast variance | +4.5% |
About half of the miss is Checkout's: the trend expected roughly February's $19,500, but orders jumped 26% and its cost reached $22,000. A driver-based forecast, expected orders × cost per order, would have caught it; that's where forecasting meets unit economics.
Why cost forecasting matters
Budgets, commitment purchases and finance plans all start from a forecast. Forecast too low and teams blow through their budgets; too high and money sits idle in the plan, or in commitments sized for usage that never arrives.
FinOps Foundation guidance has suggested variance targets that tighten as a practice matures, such as 20% at crawl, 15% at walk and 12% at run. The number to watch is your own variance, month by month.
How forecasts are built
- Trend-based. A model extends past cost forward. It's what the cloud consoles do, and it works well for steady workloads.
- Driver-based. Cost is forecast from business drivers, such as expected orders × cost per order. It catches growth and launches that a trend line misses.
- Known changes. Add what the data cannot know: migrations, new regions, price changes and commitments that expire or renew.
Common mistakes
- Trusting the trend through a step change. A migration or launch breaks the history the model learned from. Adjust the forecast by hand for known changes.
- Forecasting only the total. A company-wide number hides which team will overshoot. Forecast at the level you budget and allocate.
- Forgetting commitments. Expiring Savings Plans or reservations raise the bill overnight. Put their end dates in the forecast.
- Never checking the forecast. Measure forecast variance every month. Without it, nobody knows whether the forecast is getting better.
Where to find it in AWS, Azure and Google Cloud
| Cloud | What it is called | Where to find it |
|---|---|---|
 AWS | Cost Explorer forecasts | Cost Explorer forecasts future cost from your past usage with an 80% prediction interval, and shows no forecast when there isn't enough history, such as an account less than one billing cycle old. AWS Budgets can alert on forecasted spend. |
 Azure | Cost analysis forecast | Cost analysis shows forecast cost as a shaded part of the chart, best read on accumulated cost. The forecast covers the total of all services, not each service, and budgets can alert on it. |
 Google Cloud | Forecasted costs in Billing reports | Cost reports forecast up to 12 months ahead with machine learning, combining actual cost to date with predicted daily costs. Forecasts are not available for invoice-based billing period ranges, and budgets can alert on forecasted spend. |
How CloudYali helps
CloudYali doesn't produce a forecast itself yet. It gives a forecast its inputs: daily cost by team, account and service across every provider, unit cost views for driver-based forecasts, and budgets that check spend against the plan every day.
Related terms
Governance & planning
Budget
A budget is a planned spending limit for a period, such as a team's monthly cloud cost, with alerts when actual or forecasted spend crosses thresholds you set, such as 80% and 100%. On AWS, Azure and Google Cloud a budget warns rather than stops spending, unless you attach actions or automation to it.
FinOps foundations
Unit economics
Unit economics is measuring cloud cost per unit of business value, such as cost per order, per customer or per API call, by dividing what a product costs by how much it delivered. It shows whether spend is growing because the business is growing or because the product is becoming less efficient.
Governance & planning
Cost anomaly
A cost anomaly is spend well outside what a service, account or team normally costs, such as a day costing several times its recent average. Anomaly detection compares each day with a baseline learned from history and alerts on the difference, so a runaway job is caught in days, not at the end of the month.
Commitments & rates
Commitment coverage
Commitment coverage is the share of commitment-eligible usage that a Reserved Instance, Savings Plan, Azure reservation or Google Cloud committed use discount paid for, measured at on-demand prices. The rest was billed at full on-demand rates, so low coverage on steady usage is money left on the table.
Billing & cost types
Amortized cost
Amortized cost spreads each commitment fee across the period it covers, so a one-year Reserved Instance or Savings Plan paid up front shows up as a share of the fee every day of the term instead of one large charge on the day you buy it.
FinOps foundations
FinOps Framework
The FinOps Framework is the FinOps Foundation's open model for running a FinOps practice. It sets out principles, personas, the Inform, Optimize and Operate phases, four domains of capabilities, scopes and a crawl, walk, run maturity scale, and practitioners update it every year.
Frequently asked questions
How accurate should a cloud cost forecast be?
FinOps Foundation guidance has suggested variance targets that tighten as a practice matures, such as 20% at crawl, 15% at walk and 12% at run. Measure yours monthly and work it down.
What is the difference between a forecast and a budget?
A forecast predicts what you will spend; a budget sets what you plan to spend. Forecasts inform budgets, and a forecast running above budget is the earliest warning you get.
What is driver-based forecasting?
Forecasting cost from business drivers, such as expected orders × cost per order, instead of from past cost alone. It catches growth and launches that a trend line misses.
