Demand forecasting is hard enough when you have years of sales data to work with. When you're launching a brand new product — one that has never been sold before — you're working with nothing. No trendline to extrapolate. No seasonality to account for. No historical velocity to build from.
And yet you still have to make a purchase order decision. You have to decide how many units to produce, how much inventory to bring in for launch, and how aggressively to stock across your channels. Get it wrong in one direction and you stock out in week two, leaving money on the table and early customers disappointed. Get it wrong in the other direction and you tie up significant cash in inventory that moves slowly, eroding the margins on a product that hasn't even proven itself yet.
New product forecasting is one of the most genuinely difficult problems in ecommerce inventory planning — and one of the most consequential. Here's a practical framework for approaching it.
Why New Product Forecasting Is Different
When you're forecasting demand for an established product, you're essentially answering the question: "Given what happened before, what's likely to happen next?" Historical data anchors the forecast. Seasonality patterns emerge from the data. You can see how the product responds to promotions, channel changes, and competitive dynamics. The forecast is still uncertain, but it's grounded.
New product forecasting requires a completely different mental model. You're not predicting what will happen based on what did happen — you're estimating what might happen based on indirect signals. The inputs are softer. The uncertainty is wider. And the consequences of being wrong are often higher, because you're committing capital to a product before you know whether customers will actually want it.
The good news is that "no sales history" doesn't mean "no information." There's usually quite a bit you can learn from analogous products, market data, and early demand signals — if you know where to look.
Step 1: Find Your Analog Products
The most reliable starting point for new product forecasting is finding products that are as similar as possible to what you're launching — and using their performance as a proxy.
Analog products can come from a few different places:
Within your own catalog. If you're launching a new colorway of an existing product, a new size variant, or a product in a category you already sell in, your existing SKUs are your best analog. Look at how similar launches performed in their first 30, 60, and 90 days. How quickly did they ramp? What was the initial sell-through rate? How did velocity change after the first promotional push?
From competitors. If a competitor sells a nearly identical product, their performance gives you a market-level signal for what demand looks like. You can get some of this data from Amazon BSR (Best Seller Rank) history tools, third-party market research platforms, or simply by observing review velocity — the rate at which a product accumulates reviews is a reasonable proxy for sales velocity.
From adjacent categories. If you're launching in a category you don't currently sell in, look for products with similar price points, target audiences, and purchase behavior. A product selling at the same price point to the same demographic through the same channels is a useful reference point even if it's not the same product.
The goal isn't to find a perfect analog — it won't exist. The goal is to find the closest available proxy and use it to set a reasonable baseline rather than starting from zero.
Step 2: Estimate Market Size and Your Realistic Share
Once you have analog data to anchor your thinking, the next step is to size the opportunity realistically and estimate what share of it you can capture.
This involves two separate estimates:
Total addressable demand. How big is the market for this type of product? Category-level data from platforms like Amazon (through tools like Jungle Scout or Helium 10), industry reports, and search volume data can give you a rough sense of total demand. This isn't about precision — it's about order of magnitude. Is this a product category where 1,000 units a month is realistic? 10,000? 100,000?
Your realistic market share at launch. A new product from a new brand with no reviews and limited marketing reach will capture a very different share of demand than a product from an established brand with a large audience and strong marketing budget. Be honest about where you're starting from. Understanding how demand planning actually works at a product level will help you think through this more rigorously.
The combination of these two estimates gives you a rough demand range — not a single number, but a plausible low, mid, and high scenario that you can use to build your initial inventory plan.
Step 3: Build Scenario-Based Forecasts
For new products, a single-point forecast is almost always misleading. The uncertainty is too wide. Instead, build three scenarios explicitly:
Conservative scenario. What does demand look like if the product launches slowly — below your expectations, with limited early traction? This scenario drives your minimum initial inventory commitment. You should always be able to fulfill the conservative scenario without tying up excessive cash.
Base scenario. What does demand look like under normal conditions, roughly in line with your analog products and market share estimate? This is your planning number for initial production and inventory positioning.
Optimistic scenario. What does demand look like if the product takes off faster than expected — driven by strong reviews, viral attention, or a promotional push that performs above expectations? This scenario informs your contingency plan: what would you need to do, and how quickly could you do it, if demand significantly exceeds your base case?
Building all three scenarios explicitly forces you to think through the full range of outcomes rather than anchoring on a single estimate that's probably wrong in one direction or the other. The right demand planning approach for a new product launch looks meaningfully different from steady-state forecasting — it needs to account for this kind of uncertainty explicitly.
Step 4: Set a Conservative Initial Buy With a Fast Reorder Plan
Given the uncertainty inherent in new product forecasting, the safest initial inventory strategy is almost always to buy conservatively and plan to reorder quickly.
This doesn't mean buying so little that you stock out in week one — that creates its own problems, particularly on Amazon where early stockouts can permanently damage your product's launch trajectory. It means buying enough to cover your conservative scenario with a reasonable buffer, and having a clear plan for how quickly you can get more inventory if demand exceeds expectations.
The key variables to think through:
Your supplier's reorder lead time. How quickly can your supplier produce and ship additional units if you need them? If your lead time is 12 weeks, your initial buy needs to account for 12 weeks of potential demand at your optimistic scenario — because if you run out, you won't be able to restock for three months. If your lead time is 3 weeks, you have much more flexibility to start lean and reorder fast.
Your sales channel's in-stock requirements. On Amazon FBA specifically, running out of stock during the launch phase can suppress your organic ranking in ways that are difficult and expensive to recover from. Build enough buffer to maintain in-stock status through your initial review-building period, even if demand comes in at the optimistic end of your range.
Your cash position. Initial inventory buys for new products are inherently speculative. The right buy quantity also depends on how much capital you're comfortable committing to an unproven product. Buying conservatively and reordering is almost always preferable to overbuying and sitting on slow-moving inventory that ties up cash you need for other things.
Step 5: Treat the First 60 Days as a Calibration Period
Once the product launches, the most important thing you can do is shift from estimation mode to observation mode as quickly as possible.
Real sales data — even just a few weeks of it — is dramatically more useful than any pre-launch estimate. The first 30 to 60 days of a product's life give you the actual demand signal you've been trying to estimate. Monitoring your inventory closely during this period lets you catch divergences from your forecast early — before a stockout or an overstock has time to compound.
A few things to watch closely in the first 60 days:
Sell-through rate. Are you selling through inventory faster or slower than your base scenario projected? This is the primary calibration signal. A sell-through rate that's 2x your forecast is an immediate signal to accelerate your reorder. A sell-through rate that's half your forecast suggests you may need to revisit your pricing, positioning, or channel strategy.
Channel mix. Is demand distributed across channels the way you expected? If Amazon is dramatically outperforming Shopify — or vice versa — you may need to rebalance your inventory positioning across channels even before you know what total demand will be.
Review velocity. On Amazon, review accumulation rate is a strong leading indicator of sustained demand. A product that's accumulating reviews quickly is likely to maintain or accelerate its velocity. A product with slow review accumulation may struggle to build organic momentum.
Use this early data to update your forecast aggressively. The pre-launch estimate was always a starting point — the real forecast begins the moment you have actual sales data to work with.
The Bottom Line
Forecasting demand for a new product without sales history will always involve more uncertainty than forecasting for established SKUs. The goal isn't to eliminate that uncertainty — it's to make informed decisions in spite of it, build contingency plans for the outcomes you can't predict, and update your forecast quickly as real data comes in.
The brands that launch new products successfully aren't the ones with the most sophisticated pre-launch forecasting models. They're the ones who start with a reasonable estimate, commit to a sensible initial inventory position, and then pay close attention to what the market is actually telling them in the first weeks after launch.
Flieber's AI-powered demand planning platform helps ecommerce brands build more accurate forecasts for both new and established products — and gives teams the real-time visibility to catch divergences early and respond before stockouts or overstock have time to compound. Start for free at flieber.com.
