Demand forecasting is estimating how much people will want a product in the future, and when. Supply chains use it to decide how much stock to hold. Advertisers use it to decide when to put money behind ads and who to aim them at, and that calls for a different kind of forecast.
What is demand forecasting, in plain terms?
A demand forecast is a best estimate of future interest in something you sell. It answers two questions: how much, and when. A bakery ordering flour for next week is forecasting demand. So is a software company deciding which month to put more behind its signup ads.
The term usually comes up in supply chain work. There the goal is to know how many units to have on hand next month, and the main input is years of your own sales records. That is a real problem with good methods behind it. An advertiser has a different problem, and borrowing the inventory approach leaves gaps.
Demand planning vs demand forecasting: what is the difference?
Demand forecasting is the estimate of future demand, and demand planning is the set of decisions you make because of that estimate.
A forecast says interest in a product will rise from May 19 to 25. Planning turns that into actions. A manufacturer might add shifts and move stock closer to buyers. An advertiser decides how much to raise spend on that product during those dates and which audience to target.
The two terms get mixed up because one follows the other so closely. It helps to keep them apart, since a good forecast is only useful if the plan built on it changes what you do.
How is supply chain forecasting different from forecasting for advertising?
Supply chain forecasting asks how many units you will sell in each period, so you neither run out nor sit on extra stock. Forecasting for advertising asks when buyer interest will rise, so your ads reach people while they are looking. The first can lean on your own history. The second has to look at events outside your own data, because that is where the change in interest starts.
| Supply chain forecasting | Forecasting for advertising | |
|---|---|---|
| Question it answers | How many units will we sell? | When will buyer interest rise? |
| Main input | Your own sales history | Events outside your own data |
| Output | A quantity for each week or month | A window with a start and a stop |
| Decision it informs | How much to make and stock | When to spend on ads and who to reach |
| Time unit | Fixed periods | Windows that open and close on their own dates |
| What it misses | Demand from outside events you have never seen before | How many units to make or stock |
A supply chain forecast treats demand as a number to meet. An advertising forecast treats demand as something that rises and falls, where the job is to be in front of buyers while it is up.
Why does your own sales history miss some demand?
Your own sales history misses demand that starts outside your company, because it can only repeat patterns you have already lived through. It is still the obvious place to start, since it records real purchases. For advertisers it has four blind spots.
It only repeats what you have already seen
History-based forecasts carry past patterns forward. If a build-up happens that you have never lived through, or happens at a different time than last year, your history has nothing to say about it.
Take hiring rising month over month. Companies shop for onboarding software when that happens, because every new hire needs accounts and paperwork set up. If hiring was flat for the two years your records cover, your history gives you no reason to expect a rise.
It mixes in your own ads
Your past sales went up partly because you advertised. A forecast built on those sales learns your old spending schedule as if it were demand. If you always pushed hard in one month, the history says that month is strong, whether buyers were looking or your ads were simply louder.
New products have no history
A product you haven't launched has no sales to learn from. Neither does an existing product moving into a new category or a new audience. For these, outside events are the only place to look, which is why product launch timing depends so much on them.
Outside demand shows up late in your own numbers
The demand most worth catching comes from events that build up and converge. Home sales climbing for a third month means new owners replacing doorbells and locks. New business registrations climbing for several months means first-time founders looking for invoicing software.
Each of these starts outside your company and reaches your sales only after it has been building for a while. By the time your own numbers show it, part of the window has already passed. Verdius spots these build-ups for you and dates the window before your own numbers move.
What should a useful demand forecast for advertisers contain?
A useful demand forecast for advertisers contains a window with a start and a stop, the events behind it, the audience to reach, a ranking against other windows and alerts when each window opens and closes. A line on a chart is hard to act on, while each of these parts leads to a decision.
A window with a start and a stop. "Demand will be higher later this year" is too loose to schedule anything against. A window from May 19 to 25 is something you can plan ads around, and the stop date tells you when to pull back.
The events behind it. Knowing that interest in onboarding software is rising because hiring is climbing tells you what your ads should say. It also lets you judge for yourself whether the build-up is still going.
Who to reach. The people inside a window are a narrower group than your usual audience. The forecast should describe them in terms ad platforms already use, such as age, income, location and interests, so you can build the audience in Google Ads, Meta, LinkedIn or TikTok without translating.
A ranking. Most products have several windows ahead. Ranking them against each other shows which ones deserve the most of your spend and which you can skip.
Alerts. Windows open and close on their own dates. A forecast that tells you when a window starts and when it ends saves you from checking every day.
Verdius builds this kind of forecast for each product: ranked windows, the events behind each one, the audience to reach and alerts when a window opens and closes. It connects to the ad platforms you already use and only reads them, so you stay in charge of the ads. You can see how the windows are found.
How do you act on a demand forecast?
You act on a demand forecast by preparing before the window opens, raising spend when it opens and pulling back when it closes.
Before the window opens, build the audience in each ad platform and prepare ads that speak to the reason people are buying. New homeowners respond to a message about their new place, and companies hiring fast respond to one about getting new staff set up. Verdius gives you that audience by age, income, location and interests, ready to set up in each platform.
Decide how much of your spend moves to the product in the window. That number is yours to set, whether it is a few hundred dollars or a large share of what you spend on ads.
When the window opens, raise spend on the product in the window. The guide on when to increase ad spend covers how to raise it once buyers start looking.
While it runs, watch cost per sale and the search terms in your ad accounts. If a higher-ranked window is coming up soon for another product, you might hold some spend back for it.
When the window closes, pull back. Leaving spend at window levels after the stop means paying to reach a shrinking pool of buyers. Verdius sends an alert when a window opens and again when it closes, so you know the day to pull back.
Afterward, compare cost per sale inside the window with the weeks around it, using your own numbers. That gives you a measured read on how the window performed for your product. There are more worked examples by product type in advertising use cases.
How do I know if demand for my product is about to rise before I raise my ad spend?
You know demand for your product is about to rise when the outside event that makes people need it has been building for several weeks or months. Start by naming that event. For a video doorbell it is people moving into new homes, and for onboarding software it is companies adding staff.
Then check the direction of the build-up. Home sales climbing for a third month, or hiring rising month over month, points to more buyers arriving soon. A single busy week is weaker evidence than a steady climb.
Your own accounts confirm the rise once it starts. Search terms in your ad accounts and visits from new people move first, and sales follow. Waiting for your own sales to rise before you raise spend usually means acting after part of the window has passed.
What software gives advertisers a demand forecast?
Verdius is software that predicts when demand for a product will rise and who to target then. It gives you the kind of forecast this guide describes, one product at a time.
For each product you get:
- Optimal windows to run ads, each with a start and a stop, ranked against each other
- The events behind each window
- The audience to reach, by age, income, location and interests
- An alert when each window opens and when it closes
Verdius connects to Google Ads, Meta, TikTok, Reddit, Snapchat, LinkedIn, Microsoft Advertising and ChatGPT Ads, plus Google Analytics and Search Console. It only reads that data, and you run the ads. Higher plans see further ahead and cover more products. The pricing page lists each plan.
Do small advertisers need demand forecasting?
Small advertisers need demand forecasting as much as large ones, because anyone who runs ads makes a timing decision, even if it is only choosing the week to boost one post. A large company with steady spend can move money between products as windows open and close. A founder with one campaign can hold it until the week buyers are looking. The forecast answers the same question for both of them: when will the people who want this product start looking for it?