Inventory Planning and Forecasting for D2C

D2C brands are systematically bad at inventory planning

Most D2C brands approach inventory planning as a gut check exercise performed monthly by whoever owns operations, based on last month's velocity and a conversation with the founder about what feels right. This approach works for a while, then it stops working, and the symptoms hit all at once. Hero SKUs go out of stock. Slow movers pile up. Working capital is stuck in the wrong places. The 3PL is storing units that will never sell at full price.

The first failure pattern is velocity based forecasting without seasonality. A brand sells five hundred units a week of a hero SKU for six months, so the planner orders at that rate. They did not decompose the velocity into trend, seasonality, and promotional lift. Q4 hits, demand doubles, the brand stocks out in week three of December, and the recovery takes eight weeks because the supplier lead time is twelve and the reorder was placed too late.

The second failure is launch forecasting by founder intuition. A new product launches. Nobody builds an analog model against the last three launches. Nobody segments demand expectation by channel. Nobody builds a reorder trigger against early sell through signal. The launch either massively overshoots (dead stock for a year) or massively undershoots (waitlist email to a million people, half of whom will never convert). Both outcomes are preventable.

The third failure is reorder point chaos. Brands with large catalogs (hundreds or thousands of SKUs) cannot manage every SKU by forecast. They need reorder points for the long tail and forecasting for the head. Most brands either try to forecast everything (impossible) or set a single reorder point logic that works for some SKUs and not others. The right architecture is segmented.

Our approach

We run inventory planning as a six step engagement that ends with a working planning system your ops and finance teams use every week.

Step one is catalog segmentation. We classify every SKU into tiers based on velocity, margin, and strategic importance. Hero SKUs get full forecasting. Mid tier SKUs get simplified forecasting with reorder point logic. Long tail SKUs get reorder point only. New launches get analog based modeling. This segmentation is the foundation everything else sits on.

Step two is demand forecasting for the head catalog. We build forecasts with trend, seasonality, and promotional decomposition. We pull two to three years of sales data where available, adjust for stockouts and promotions, and build a baseline forecast with explicit confidence intervals. Forecasts are not single point estimates. They are ranges with assumptions documented.

Step three is reorder point logic for the mid and long tail. We set reorder points based on lead time, lead time variability, demand variability, and service level target. The service level target is set by strategic importance: hero SKUs target ninety eight to ninety nine percent in stock, mid tier target ninety five percent, long tail target ninety percent. Setting the target explicitly is the key decision that unlocks everything downstream.

Step four is launch and promotion modeling. We build an analog model for new launches using the last three to five comparable launches as reference points. We also build a promotional lift model for the head catalog so promotion planning and inventory planning are the same conversation, not two separate conversations.

Step five is the S&OP cadence. We design the weekly, monthly, and quarterly review cadence. Weekly is tactical: adjust POs based on the last seven days of signal. Monthly is planning: review next ninety days of forecast against actual. Quarterly is strategic: merchandising, finance, and ops align on the next quarter's inventory posture. The cadence is documented and the meetings have clear decision rights.

Step six is reporting. We build dashboards covering in stock rate, stockout incidents, dead stock aging, inventory turns, working capital tied in inventory, and forecast accuracy. Forecast accuracy is the single most important metric because it tells you whether the planning system is getting better over time.

What you get

▸ A segmented catalog with SKU classification by tier and strategic importance ▸ A demand forecasting model for the head catalog with seasonality decomposition ▸ Reorder point logic for mid and long tail SKUs with explicit service level targets ▸ A launch analog model and a promotional lift model ▸ A documented S&OP cadence covering weekly, monthly, and quarterly reviews ▸ Dashboards covering in stock, stockout, dead stock, turns, and forecast accuracy ▸ Integration with your existing ERP or inventory system ▸ A written playbook for the planning team ▸ A ninety day tuning review with forecast accuracy against actual

Timeline

Weeks one and two are catalog segmentation and data preparation. Weeks three and four are forecasting model build and reorder point logic. Week five is launch and promotional modeling. Week six is S&OP cadence design and dashboard build. Week seven is training and go live. Week eight is stabilization. Ninety days later we run the forecast accuracy review.

Mini case anatomy

A composite from a growth stage D2C personal care brand with around four hundred SKUs. They ran planning on spreadsheets. Stockouts on their top ten hero SKUs were common, especially in Q4. Dead stock was a growing line item on the finance review. Working capital tied up in inventory had grown faster than revenue for two consecutive quarters.

We segmented the catalog. Ten hero SKUs drove about half of revenue. Fifty mid tier SKUs drove another third. The remaining three hundred and forty SKUs drove the final sixth. The existing approach treated all of them roughly the same.

We built a forecasting model for the hero ten with full seasonality decomposition. We set reorder points for the mid fifty with a ninety five percent service level target. We set reorder points for the long tail with a ninety percent service level target. We built an analog model using the last four launches as reference and applied it to the three launches scheduled in the next quarter.

We designed the S&OP cadence with weekly tactical reviews between ops and finance, monthly cross functional reviews including merchandising, and a quarterly strategic review with the founder. Dashboards went live in week seven.

Ninety days after launch, stockouts on hero SKUs dropped to near zero. Dead stock aging stabilized and began to improve as the long tail reorder logic stopped replenishing slow movers at the prior rate. Working capital tied up in inventory stabilized relative to revenue. Forecast accuracy on the hero SKUs came in at a level that made finance comfortable building reliable cash models for the first time.

The planning system stopped being a source of anxiety and became a source of confidence. Merchandising got real lead time on launch planning. Finance got reliable cash forecasts. Ops stopped getting the 3am message about a stockout.

Related services and reading

Inventory planning pairs with 3PL selection, order management systems, and fulfillment audit. Brands also rethinking returns should look at returns program, since return timing affects sellable inventory.

On the customer experience side, accurate inventory data powers a trustworthy post purchase UX and reliable AI support agent setup. Recommended reading: ecommerce customer lifetime value and post purchase experience and repeat buyers. Parent hub: ecommerce operations.

FAQs