Inventory Optimisation
Lightspeed, 2024
1
Problem Discovery
Problem statement
Merchants who regularly replenish stock rely heavily on past sales data to predict future demand. However, this method does not consider the periods when products were out of stock.
As a result, sales predictions become inaccurate, leading to under-ordering and missed sales opportunities. Merchants are left with two difficult options: manually conducting time-consuming workarounds or risking inventory shortfalls.
Above: Example use case of underordering due to stock out periods
How might we help merchants take into account past stock-out periods when ordering so that they can more accurately order and not miss a sale?
User research
To gain a deeper understanding of the pain points merchants face, we conducted user interviews with retail and inventory managers – the professionals responsible for stock ordering. Through these interviews, I mapped the user journey to uncover key challenges within the
current process.​
Time-consuming workarounds
User goals
​Based on the insights from user interviews, we identified the pain points in the journey and turned them into user goals which informed the solution:
​
-
Reduce the need for merchants to cross-reference multiple places to make informed decisions
-
Help merchants easily identify and understand when products experience stockout periods
-
Provide a more accurate representation of a product’s historical performance by factoring in stockout periods
-
Build trust in the insights provided, so merchants can confidently rely on the system for ordering decisions​
2
Solution Discovery
Defining "Stockout"
We first needed a clear, shared understanding of what counted as a
stockout period.
We defined stockout as a day with no inventory and no sales. This distinction was important—since there could be days with no sales, but that didn’t necessarily mean the product was out of stock. Then using this definition, we developed the new forecast calculation.
New intended journey
​Then, having the user goals in mind, I mapped the ideal user journey in words. This will be the foundation of the solution moving forward.
Lo-fi wireframing
Once the journey was clear, I moved into wireframing to help bring the user journey into the product’s context, exploring multiple directions.
Explorations
Purchase Order flow
This is where merchants go once they’re ready to place an order. But during that time, there were limited metrics or insights that were available to the user on the page. And that was a key consideration for me since each merchant is different – they use different metrics based on what they need or the context of their business. And this page just lacked that broader context.
Dedicated Forecasting page
We quickly realized that this will require users to adapt to a new workflow but also adds redundancy, since most of the functionality that we’ll need in this new page already exists somewhere else.
Granular sales data
Giving users a more granular breakdown of sales trends where they can see sales on a weekly or daily basis. But this felt like it was optimizing for a different problem focusing more on trend analysis. From a technical point of view, this required significant additional development time.
Chosen solution
Enhancing existing reports with new metrics
The final direction we landed on was enhancing the existing report builds by providing new metrics. And there are a couple of benefits to this:
-
Scalability – as we build on top of reporting tools, the forecast
workflow will benefit from it. -
Familiarity – reduces friction by leveraging an existing workflow.
-
Flexibility – Merchants could mix and match metrics based on
their needs. -
Better real estate – The reports page had more space for
presenting insights.
Integrated Data Dashboard
I designed an intuitive, unified report that consolidates sales data and stockout information into a single view.
This eliminates the need for merchants to manually combine data from separate reports.
Forecasted Demand and
Suggested Order Quantity
We implemented an algorithm that adjusts historical sales data to account for stockout periods.​ Additionaly, users can modify the duration for how long they want to forecast, and the algorithm calculates the suggested order quantity accordingly.
The result is a more accurate forecast of future demand, reflecting real sales trends instead of artificially inflated or deflated data due to stockouts.
Missed Sales & Stockout Periods
I introduced a column that clearly states how many missed sales there were due to stockouts and allows users to understand the duration of the stockout period. I used progressive disclosure to give more context, providing a deeper view of the stockout's impact on sales.
Transparent Calculation Model
To build trust, an easy-to-understand explanation of how the sales forecasting model works, outlining how stockouts are factored in. This helps merchants feel confident in the insights generated.
Giving Users Control
I provided users with the flexibility to toggle these new metrics on and off, ensuring the reports remain adaptable to both detailed forecasting and more straightforward reporting needs.
3
Impact
User outcome
By implementing the above solutions, merchants no longer need to choose between time-consuming workarounds and inaccurate sales predictions. They now have a reliable tool that provides them with:
​
✅ Unified Reports: No more toggling between different reports. Merchants can now view both sales and stockout data in a single, easy-to-navigate dashboard.
​
✅ Stockout Awareness: Merchants can quickly identify and understand stockout periods, which enables them to adjust their ordering based on a more accurate view of product performance.
​
✅ Accurate Replenishment Forecasting: Sales data is adjusted to account for stockouts, offering merchants a clearer, more accurate representation of demand, ensuring they don’t miss sales.
✅ Confidence in Data: Clear, transparent calculations help merchants trust the system and make data-driven decisions with confidence.
Business outcome
Enhanced Marketability: This feature enabled the product to meet the demands of mid-market merchants with large-scale, complex needs, strengthening its appeal in a competitive market. Sales teams leveraged these improvements to better position the product as a robust solution for high-demand environments, leading to an increase in Average Revenue Per User (ARPU).