How data apps helped Huckberry save $1M+ in inventory planning

How do you forecast demand across 100,000 SKUs — and actually get it right?

That’s the puzzle Huckberry’s small but mighty data team was tasked with solving. On one side, the marketing team was pushing for growth, ramping up demand for seasonal collections, core products, and limited drops. On the other, the planning team had to ensure the right inventory was on hand — not too much, not too little — across hundreds of thousands of product variants.

Miss the mark, and the company risked stockouts, lost revenue, or bloated inventory that tied up cash. But legacy BI tools couldn’t get granular enough to answer the real questions: What’s the right inventory spread across sizes and colors? What happens if we shift a product promotion? Can we trust the forecast for a brand-new SKU?

That’s where Huckberrry’s Head of Data, Ben Saxe, and his team stepped in.

Rather than relying on expensive, rigid forecasting tools — or endlessly reacting to ad hoc requests — Huckberry’s data team built something better. In just weeks, they delivered an interactive forecasting app that gives business teams the power to run scenario modeling, forecast at the SKU level, and zero in on what matters most: profit.

The inventory balancing dilemma: Keeping products in stock without overstocking

The popular “one-stop men’s shop” has over 100,000 SKUs (stock keeping units) between clothes, gears, and gifts across the entire website. The constant challenge is to keep the revenue-driving items in stock while managing the inventory of new and seasonal products. Ordering too much or too little of any SKU poses risks — twisting the inventory dial too far in either direction can create problems:

• If Huckberry allows some sizes or colors of their main revenue-drivers to run out of stock, then they leave money on the table.

• If Huckberry buys too much stock, then the company’s money will be tied up in products that it can’t sell fast enough, restricting it from spending that money elsewhere.

It’s a tough balance that many retailers struggle to strike. To account for any imbalances, some retailers buy more stock than they anticipate selling and then discount the remaining batch that they couldn’t sell at full price. Others buy only the stock they anticipate selling, knowing that they might run out and leave some customers disappointed.

Huckberry, however, is primarily a full-price brand — they want to keep markdowns to a minimum across the company, especially with their best sellers.

To forecast and plan with precision, Saxe’s team built a data app that helped the company make gains without leaving profit on the table.

New precision for inventory planning, delivered by the data team

The data team built two data apps in Hex that give marketing and planning teams the ability to predict future demand with far more accuracy than spreadsheets could provide. These apps help them keep popular items in stock, maintain healthy inventory turnover, and plan for seasonal demand.

"Doing all of this for 100,000 SKUs with a three-person planning team is a lot to manage," Ben said. "We took cumbersome processes and added a layer of sophistication, which created tremendous value for us."

One major improvement: eliminating blind spots. When a product is new and lacks historical data, teams can use seasonality curves from similar products. They can also adjust for holidays to avoid overstocking or stockouts caused by seasonal spikes.

Unlock 1: Real-time inventory health, now and weeks ahead

With one app, the planning team can instantly see how their supply is tracking for any product — and especially for the next four weeks. They can filter by units, normalize demand, and immediately understand: How did last week go? What does the next month look like?

If they want a short-term view, it’s easy to compare current inventory to recent weeks and see at a glance what’s available to sell versus what’s on order. This enables accurate predictions up to 26 weeks ahead, helping them fine-tune stock levels and avoid under- or overstock.

Unlock 2: Recommendations for color and size SKUs

Once they feel good about their forecast, they can take the results and move it into another Hex app to get a recommendation of what the should look like across colors and sizes.

The planning team can adjust the model and update forecasts in real-time as trends change. For example, if the product team considers dropping an inseam length from a popular pant, planners can instantly see the impact on the forecast.

“If you change this factor, or your confidence, or the amount that you want the model to swing, we can give those as input variables.” Ben said. With Hex, the team can get as granular as needed. “Hex takes this to the nth degree,” Saxe adds. “You’re not just filling in fields — you’re interacting with the model.”

Now, the planning team can:

• Predict which SKUs need replenishing

• Forecast what needs to be marked down or more aggressively marketed

• Plan for scenarios at different price points or size mixes

Data app vs. dashboard

The data team had anticipated a few months of work to build out a smarter forecasting solution, but with Hex, were able to build apps that provided clearer forecasts in just a few weeks. To test the improvement, they compared Hex’s forecast model against the previous approach, which Saxe describes as a “clunky spreadsheet process.” In one case, Hex’s model recommended ordering half as much inventory, freeing up significant cash flow over time.

Additionally, before Hex, fulfilling ad hoc requests was a slow, manual process. “I’d have to write a ton of SQL, modify the BI tool, add fields, build the report, and create a dashboard,” Saxe said. With every new request, the process started over. “It was extremely cumbersome.” But now, with Hex, business users can make adjustments themselves — no bottlenecks, no silos.

“What's so incredible is that when business users know something we don't, they can actually adjust it themselves,” Saxe said.

Instead of purchasing expensive solutions, Huckberry’s lean data team used Hex to build more accurate, collaborative tools that keep the data team’s work connected and scalable.

The payoff: Over $1M+ saved, hours freed, and better decisions

Before Hex, inventory planning was an extremely cumbersome process and less accurate overall — with either a bit more core products unsold than what would be ideal, or popular SKUs running out mid-season. Now, Huckberry runs a leaner, sharper core replenishment model, which has unlocked over $1M in savings.

"It really pays off in dividends to accurately forecast and look to the future for some of our big product campaigns," Ben said. "We reviewed this with the CFO and he is advertising this to the rest of the company as 'a million-dollar product.'"

Alongside the financial returns, the planning team is now saving 4 hours a week and strengthening its relationships with suppliers. Better projections eliminate the need for constant seasonal renegotiations, leading to smoother communication and fewer adjustments. And, in addition to getting time back, the planning team is able to deliver more accurate results for the fringe items.

The bottom line? With Hex, Huckberry transformed inventory planning from a high-stakes guessing game to a data-driven advantage — freeing up time, capital, and confidence to focus on what’s next.