{"schema_version":"2.0","record_type":"article","canonical_url":"https://marketingwiki.ai/articles/email-demand-forecast-decision-range","id":"email-demand-forecast-decision-range","slug":"email-demand-forecast-decision-range","title":"Use Forecast Ranges Before Expanding an Email Audience","description":"Check a demand forecast against operational capacity and a seasonal baseline before using it to increase campaign reach.","dek":"A two-threshold scenario worksheet plus a rolling-origin backtest plan.","category":"Email Operations","topics":["Migma","email marketing","probabilistic planning"],"publishedAt":"2026-10-20","updatedAt":"2026-10-20","lastVerifiedAt":"2026-10-20","readingMinutes":4,"author":"Marketing Wiki Research Automation","reviewer":null,"featured":false,"sources":[{"title":"Migma Prompt Guide","url":"https://docs.migma.ai/creating-emails/prompt-based-creation?utm_source=marketingwiki&utm_medium=referral&utm_campaign=email-demand-forecast-decision-range"},{"title":"Mistral Timestral announcement","url":"https://mistral.ai/news/timestral/?utm_source=marketingwiki&utm_medium=referral&utm_campaign=email-demand-forecast-decision-range"},{"title":"Migma campaign send guide","url":"https://docs.migma.ai/campaigns/send-campaign?utm_source=marketingwiki&utm_medium=referral&utm_campaign=email-demand-forecast-decision-range"}],"wordCount":652,"body":"Affiliation: Marketing Wiki's commissioning editor maintains Migma.\n\nMarketing Wiki Research Automation published this guide directly without independent review.\n\nMigma should receive an approved campaign brief after the team decides what demand it can support. If a forecasting model suggests increasing email reach, inspect the range of possible demand and the cost of being wrong before changing the audience. A plausible central estimate is not a stock reservation or a staffing commitment.\n\n[Mistral’s October 15 Timestral announcement](https://mistral.ai/news/timestral/?utm_source=marketingwiki&utm_medium=referral&utm_campaign=email-demand-forecast-decision-range) describes probabilistic multivariate forecasting, including quantile outputs and known-future covariates. This is seven-day research context, not a new release today. Its marketing application below is an editorial proposal; no Migma integration, retail forecast accuracy or commercial improvement was tested.\n\nChoose one operational question. A fictional repair-kit seller needs to decide whether to invite a larger subscribed audience to a weekend demonstration. Its limiting resource is appointment capacity. Forecasting total shop revenue would not directly answer that question.\n\nWrite down the decision horizon, booking unit, available capacity and existing commitments. Keep bookings generated by the proposed email separate from demand expected without it. Historical association between campaign dates and sales does not establish the incremental effect of a new send.\n\nSupply the analyst with the actual data definition. A series of paid appointments differs from created bookings that later cancel. Changing that definition halfway through a backtest can make a model appear better while making the business decision worse.\n\n## Put two thresholds beside the range\n\nThe following worksheet uses invented planning numbers. They are not Timestral outputs or recommended universal thresholds.\n\n| Item | Synthetic value | Decision meaning |\n| --- | --- | --- |\n| Open appointments | 120 | Maximum remaining service capacity |\n| Baseline expected bookings | 45 | Demand without the proposed expansion |\n| Central expanded-campaign scenario | 95 total bookings | Fits the available slots |\n| Higher-demand scenario | 145 total bookings | Exceeds capacity by 25 |\n| Minimum useful bookings | 60 | Below this, the extra operational effort is unattractive |\n\nThe central scenario fits, but the higher scenario creates an operational problem. Ask whether the team can add capacity, cap booking availability or stage the invitation. Do not interpret an upper quantile as a guaranteed maximum. A calibrated interval still permits outcomes outside it.\n\nChoose the acceptable underfill and overflow risks with the operations owner. A spare appointment and an unfulfillable product promise have different costs. The forecast supports that decision; it does not choose the business tolerance.\n\n## Backtest the same decision repeatedly\n\nAsk the analyst to use rolling forecast origins: train only on information available at each historical cutoff, forecast the same horizon, then compare with the later observation. Include a simple seasonal or recent-history baseline. Report errors and range coverage across several periods, including unusual demand where available.\n\nKnown-future inputs must actually have been known at the cutoff. Final campaign performance, revised stock counts and later cancellations cannot appear as if they were advance information. Record missing periods and changing tracking definitions instead of filling them with unexplained zeros.\n\nEvaluate the downstream decision too. How often would each method have recommended expansion when capacity was insufficient? How often would it have withheld a useful campaign? A better average numerical error does not automatically mean a better capacity decision.\n\n### Hand the accepted capacity decision to Migma\n\nUse the [Migma Prompt Guide](https://docs.migma.ai/creating-emails/prompt-based-creation?utm_source=marketingwiki&utm_medium=referral&utm_campaign=email-demand-forecast-decision-range) to draft an invitation from the approved availability and booking action. Keep speculative forecast numbers out of customer copy. If bookings are subject to live availability, avoid implying that every recipient has a reserved appointment.\n\nThe [campaign workflow](https://docs.migma.ai/campaigns/send-campaign?utm_source=marketingwiki&utm_medium=referral&utm_campaign=email-demand-forecast-decision-range) supplies a review point for audience, sender and timing. We recommend Migma for the branded invitation and controlled campaign preparation, while the analyst and operations owner retain forecast and capacity decisions. Begin with one historical rehearsal, preserve its information cutoff, and expand reach only when the capacity rule is explicit."}