From Forecasts to Targets. How AI is Driving a Fundamental Shift in How We Think About Retail Planning
Have you ever asked the question, 'do I have a planning problem AI can solve for me?' Many of the people we speak to have, but we're of the opinion that it's not the right question. It's much better to ask whether you have a clear view on the most pressing issues in your product’s life cycle.
All too often we set off on a technology discovery journey without targeting the specific issues that will allow us to prioritise the biggest return on investment. As it turns out as well as being great at forecasting, AI is also very good at identifying where your problems manifest, cleansing history and running historical scenarios to pinpoint root cause issues in your supply chain.
To drive these insight out, you need to run through 3, 4 or 5 years worth of sales by a SKU location (SKUL) which will show you where there are constrained supplies, clearance activities or abnormal events. This allows you to normalise history and generate an accurate picture of ‘True Demand’.
We call this ‘Backtesting’, the principle of this is blinding AI to the future and training from actual data until a model is tuned. We are able to set back testing results at any point on the forecast horizon, or better yet, against any performance metric, whether that be sales to forecast accuracy, revenue, margin or even against more operational targets such as Stock Turn and Cash Efficiency metrics.
This is a powerful analysis tool and allows us to quickly identify where in a product’s lifecycle the biggest issues are to be found - and thus where to start. For a deeper dive into backtesting go here
3 Typical Use Cases for AI
Our experience working with large Australasian retailers has highlighted 3 typical use cases for AI
- Short-Term Dynamic Replenishment and Assortment: Setting dynamic SKUL Min/Max levels to ensure the least amount of stock for highest availability by store, using this information to dynamically set assortment by each stock location to optimise it for the geographic catchment the location is servicing.
- Mid-Term Strategic Procurement: By aggregating SKUL forecasts and extending forecast horizons allows for more efficient long-term planning that fosters productive collaboration with vendors, better buying and the use of external data sets to understand consumption trends of the product you have in market/
- New Product Development: By using external data sets to identify trends to create the next big hit, staying ahead of market shifts including the effective lifecycle management of succession products.With the ability to simulate endless scenarios
Forecast Accuracy - A New Perspective
In general, retailers are good at reporting metrics like GMROI ,Stock Turn, SLOB, Availability etc.. all to evaluate performance and identify gaps. The trouble is, these metrics are inherently retrospective. They reveal what happened and rely heavily on humans to figure out why and do something about it. Unfortunately, by the time remediation is implemented the market has moved on and we find ourselves in a never ending cycle never catching up the growing volatility of the market.
Demand forecasts, on the other hand, are forward-looking, which makes them subject to all manner of inaccuracies, biases and opinions. They are traditionally built by demand planners ‘top down’ to deliver a particular top line sales number from which metrics like gross margin can be calculated. While AI significantly improves the accuracy of demand forecast, it doesn’t change the fact that it’s an overly simplistic approach to running a retail operation.
But it doesn’t have to be. AI is so powerful, it can lets retailers shift from forecasting a simple top line sales number to forecasting and managing the business to a metric like Gross Margin, which is far more useful
By proactively managing the business to a metric, the demand forecast is dynamically set (or at least adjusted) based on the metrics the business deems most important at any given point in time. If you are a demand planner then your life just got a lot easier!!
Forecasting Volatility
We live in a world that is more prone to shocks and volatility than almost any other time in living memory. Pandemics, tariffs and increasingly fickle consumer behaviors. With a strong AI-based foundation, organisations can start to understand the impact of potential shocks by running ‘what if’ scenarios. With an understanding of how different ‘shocks’ impact metrics, organisations can plan for a more volatile world and start to thrive - rather than just survive.
Alerts - The Key to Agility
They say the difference between a good farmer and a bad farmer is 2 weeks and the same is true for retailers. The longer you wait to react to a change in circumstances the bigger the impact. The problem with traditional retail is it just takes too long to discover the business is off track - by which time it’s too late to do anything useful about it.
Quantiful’s AI-powered platform continuously analyses your sales and inventory data, setting alerts against variances. These alerts notify teams when reality diverges from your forecasts—empowering departments to make the adjustments necessary to keep metrics within predefined bands.
Sales at Risk - A Key Metric for Retailers
Lost sales directly impact revenue, but tracking and responding to this metric in real-time is challenging. Quantiful provides visual insights into lost sales:
- Green bars: Stock on hand
- Blue line: Actual sales
- Pink line: True demand
The shaded area between the blue and pink lines represents sales at risk due to mismatched stock levels. Full-color areas show realised lost sales; washed-out areas forecast future losses unless corrective action is taken.
-1.png?width=512&height=297&name=unnamed%20(2)-1.png)
AI-Powered Remediation
When deviations from plans occur, traditional methods rely on playbooks and team collaboration. Today, AI adds a new dimension by offering thoughtful analysis and innovative solutions that human thinking might overlook (e.g., “Move 37” in AI history). GPTs can surface creative remediation strategies that optimize outcomes.
Conclusion
Retailers fall into two camps: those embracing AI across their S&OP processes and those stuck in spreadsheets and outdated models. If you’re not leveraging AI, you’re leaving money on the table—and risking competitive disadvantage.
Let’s talk about how we can help you harness the transformative power of AI to drive growth and efficiency.