Navigating the perfect storm: Profitable pricing decisions with data and AI

Stage one: Understanding the role of pricing in your business.

Before making AI-driven pricing decisions, you need to fully understand the role of price within your business. Why? To uncover the ‘size of the prize’, set out the roadmap of activity and tools to get there, and get executive level buy in.

Our approach.

1. We’ll start by assessing your existing approach to pricing. We do this by discussing the challenges, risks, opportunities and priorities with your pricing stakeholders. This ensures price adjustments align with customer expectations (e.g. value-oriented customers expect low prices).
2. Then we analyse your data using intelligent KPIs* (key performance indicators) and price sensitivity models.
3. These models recommend customer and data-driven adjustments, to optimise margin and volume across your product assortment. Recommendations are based on customer behaviour and intelligent KPIs, so we can forecast the likely impact of price moves.
4. We group your existing product hierarchy based on price points and tiers, to help you understand customer, promotion and sales participation.
5. We present BI and actionable insights. This helps your trading, merchandising and finance teams identify drivers of sales performance by price band and range, plus opportunities for growth.
6. Finally, we outline the size of the prize for executive alignment. The financial benefit that could be realised by using data and AI to optimise pricing decisions.

*Our data-driven approach uses intelligent KPIs to identify range gaps and under/over investment in price and promotion:

• Price elasticity – What will happen to volume if I change price?
• Visits and conversion – Are people seeing the product? Do they buy it when they see it?
• Stock fragmentation – Do I stock the size/colour/flavour people want to buy?
• Price gap vs competition – Am I competitive on price or out of line?
• Price change history – How many times has the product’s price changed? What was the sales volume at different price points?

As part of stage one, we carry out a strategic analysis and make key recommendations aligned to the following outcomes:

1. Which categories are you over promoting in and where can you promote more?
2. Where does value matter the most and where is competitive pricing vital?
3. Where can base price be inflated to unlock margin for reinvestment? Where do we need to invest in core pricing?
4. Where are the opportunities to stretch the price architecture with more premium ranges/tiers?
5. Identify headroom by range and category and deliver recommended selling prices.

What we deliver.

At the end of the project, we’ll deliver a presentation outlining our pricing recommendations. We’ll arm you with insight and practical actions for improved sales and margin.

Stage two: Pricing decision intelligence.

Using stage one insights, it’s time to leverage AI-driven recommendations for daily or weekly price actions.

We combine AI with visual BI to empower traders, merchandisers and finance teams to:

• Quickly identify over and under-performing products.
• Diagnose performance issues.
• Make corrective actions, backed by evidence and data.
• Measure the impact of their decisions.

Our approach.

1. Our AI platform automatically delivers insights and recommends actions every Monday morning, maximising efficiency. These are delivered as highly visual and easy to consume BI. It includes intelligent KPIs such as price elasticity, visits, conversion, stock fragmentation, price gap to competitors and price change history.
2. We work with you to develop a data-driven framework, so you can focus on the product lines with the biggest potential P&L impact – whilst minimising risks within your customer base.
3. Then, we’ll work with you to create sales forecasts based on our strategic recommendations, so you can monitor and measure your performance versus expectation. This allows for fast course correction if sales volume is too high (impacting availability) or if volume and margin is lower than anticipated.
4. Our decision intelligence framework reviews all products across your assortment and identifies the following actions:
   • Base price down/up.
   • Promote more/less.
   • Promote at existing price.
   • Review image/description.
   • Markdown/clear stock.

There are major benefits to a decision intelligence approach:

• Speed to insight: Traders, merchandisers and finance teams can access key action lines when the working week begins.
• Data-driven margin improvements: Informed by intelligent KPIs and price sensitivity, not just volume and margin.
• Solve problems: Identify whether price is the reason for over/under performance versus your forecast.
• Fast course correction: Track performance to understand which products have not responded to price changes as expected. Where do you need to quickly inflate or invest to protect margin and volume?

What we deliver.

Expect daily/weekly price actions and intelligent KPIs in the form of data feeds and visual BI (charts, tables and text). Just let us know the best format and frequency for your business. Measurement is ongoing, so you can course correct quickly.

Stage three: Data-driven pricing optimisation.

We’ll take the data from stage two to inform data-driven price moves across your entire assortment. As before, you’ll be empowered with timely, actionable insight every Monday morning – and throughout the week.

Let’s start with a key question: What’s the optimal price given products will cannibalise each other’s sales and fight for share of demand? The answer often relies on the art of experience, rather than AI-driven recommendations. That’s where we come in.

Our approach.

1. We train our pricing models to meet your objectives, using a combination of price elasticity, sell-through rates and demand transference. You might be aiming to sell through at maximum margin or maximise value for a specific promotion.
2. Next, we leverage machine learning to predict the impact of price changes. We’ll recommend optimum prices and forecast the impact.
3. We go beyond traditional methods by using:

• • Need states and demand transfer groups. Grouping products into need states (based on how customers shop) means we can forecast the impact of one product’s price change on another. For example, if we change the price of a black dress, demand for similar black dresses will increase or decrease as a result.
  • Cannibalisation and demand transference. Our algorithms look at how demand changes according to product interactions within a need state. This means we can determine the substitutability of product pairs and understand the impact of one demand change on another. A critical understanding for significant price and promotion investments.
  • Price optimisation. How does price affect product sales (i.e. price sensitivity)? How does demand transfer between products? We use historical data combined with AI to predict how a price change impacts sales volume.
  • Business constraints and guardrails. Every retail business operates with subtly different pricing strategies. Our solution is flexible and doesn’t rely on a one size fits all approach.

We use both standard and bespoke guardrails to ensure any price recommendation is compliant and aligned to your business rules. This includes price psychology (e.g. 99 vs round price points), hierarchy (e.g. premium tier 20% more expensive), price gap (e.g. matched or cheaper than competitors) and compliance (e.g. no price moves during price establishment).

The outcome is simple: recommended prices to meet your objectives. And the result? Increased margins, availability and sell-through.

What we deliver.

You’ll receive a cloud-based, price modeller software application. With a highly visual, user-friendly interface, it’s designed to be used by pricing decision makers.

Traders, merchandisers and finance teams can model the impact of multiple price moves and see the outcome at line level, plus total category performance. Most importantly, users can override price recommendations (for example, if they’re unable to execute against the volumes predicted e.g. due to delayed shipments or supplier agreements.)

Total business forecasts are reported on in a BI suite, so you can course correct on any lines not meeting expected volumes.