> ## Documentation Index > Fetch the complete documentation index at: https://docs.topsort.com/llms.txt > Use this file to discover all available pages before exploring further. # Campaign Forecasting Campaign Forecasting estimates how a campaign is likely to perform before it goes live. It turns a campaign setup into a projection of expected results, so advertisers and ad operations teams can plan spend, set expectations, and compare configurations without having to launch and wait. ## What a Forecast Tells You For a given campaign setup, a forecast projects the outcomes that matter for planning: how often the campaign is expected to be shown, how much engagement it is expected to generate, how much of the budget it is expected to use, and the sales it is expected to drive. Every projection is returned as a range rather than a single number. The range reflects how much natural variation to expect for a campaign of that size and duration, and it is the part of the answer that should drive planning decisions. ## What the Forecast Takes Into Account A forecast is built from three kinds of context. No advertiser needs to supply historical data, because everything beyond the campaign setup itself comes from the marketplace's own activity. ### The Campaign Setup What is being promoted, where it is eligible to appear, when the campaign runs, how much it can spend, and which bidding strategy and performance goal it uses. ### Recent Marketplace Performance How the marketplace as a whole has been behaving lately: the level of shopper activity, how engaged shoppers are with ads, and the prevailing cost of advertising on it. This is what makes a forecast specific to each marketplace instead of a generic industry benchmark. ### Demand on the Targeted Inventory How much shopper attention the campaign's specific targets attract, across the products, searches, categories, or placements it is set up around, and how contested that inventory is by other advertisers. ## Supported Campaigns Forecasting is available for sponsored listings, banner, and sponsored brand campaigns, under both automated and manual bidding. The projected metrics offered depend on the format, since not every format is measured the same way. ## Where Forecasts Appear Forecasts are shown in the campaign creation flow of the ad platform, so advertisers see expected results as they configure a campaign and can adjust budget, dates, or targeting and see the projection update. They are also available as a standalone component of [Ad Intelligence](https://docs.topsort.com/en/knowledge-base/ad-intelligence.md), so marketplaces can surface projections inside their own planning tools and workflows. ## How to Use a Forecast Well > **Plan With the Range** > > The range is the forecast. Planning against the single midpoint value > overstates how precise any pre-launch estimate can be. > **Forecast One Campaign at a Time** > > Each forecast describes a single campaign configuration. Adding forecasts > together does not describe what a set of campaigns would do collectively. > **Re-Run After Changes** > > Budget, dates, targeting, and bidding strategy all move the projection. > Forecast again after editing a campaign rather than reusing an earlier > estimate. > **Keep the Horizon Short** > > Forecasts are built for near-term planning, roughly up to a month ahead. > Longer horizons are better handled as a sequence of shorter plans. ## What Forecasting Is Not A forecast is an estimate of likely performance, not a commitment. Requesting a forecast does not reserve inventory, guarantee delivery, or book a campaign into a slot. For campaigns that go through the auction, actual results also depend on live shopper behavior and on how other advertisers bid for the same inventory while the campaign is running. A pre-launch estimate reads that competition from recent history, so a change in it during the campaign period is not something the forecast can anticipate. Forecasts are most reliable where the marketplace already has meaningful activity on the inventory being targeted. Brand-new inventory, newly launched formats, and targets with very little history produce wider ranges, which correctly signals that less is known. *** **Last updated:**  This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.