Are Discounts Killing Your Profit? A 4-Step AI Fix
Introduction
Are you sure every sale you make is actually making you money? Many businesses focus on boosting sales, but they don't realize that some products, especially heavily discounted ones, can secretly drain their profits. This guide shows you how to use Terno AI to find these problem products and build a smarter, more profitable strategy in 5 easy steps.[See the full Terno AI conversation here]
To explore this, we have used Terno AI to analyze the Flavors of Cacao dataset. The goal was simple: to map the global cocoa supply chain and highlight opportunities for chocolate companies, retailers, and even investors.[See the full Terno AI conversation here]
Prompt: Give me a summary of this dataset. Also, visualize the distribution of Sales, Profit, and Discounts



- What it is: This visual shows two histograms: "Sales Distribution" and "Profit Distribution".
- What it shows:
- The Sales Distribution chart shows that the vast majority of sales transactions are low-value, clustered on the far left (under 5k).
- The Profit Distribution chart shows that most transactions are clustered right around $0 profit, but there's a wide spread, with significant instances of both large profits (up to 5,000) and large losses (down to -5,000).
- Purpose: This gives the a quick "lay of the land." It establishes that while most sales are small, the profit and loss on individual sales can vary dramatically, setting the stage for investigating why.
Step 1: Find Which Products Are Losing You Money
First, you need to know which items are your winners and which are your losers. Instead of digging through spreadsheets, an AI tool can give you the answer in seconds.
- Look at the big picture: Start by asking Terno to show you sales vs. profit for each product category. You might be surprised to find that some of your best-selling categories are barely profitable.
Prompt: Visualize sales vs profit by category and sub-category.


– What it is: This is a set of three scatter plots, one for each main product category: Furniture, Office Supplies, and Technology.
– What it shows: Each dot on the charts is a product sub-category. Its horizontal position (X-axis) shows its total sales, and its vertical position (Y-axis) shows its total profit.
– Purpose: This is the first step in finding the "problem products". Readers can instantly see that some sub-categories (especially in the 'Furniture' chart ) generate high sales (are far to the right) but are below the '0' profit line, meaning they are losing money.
- Pinpoint the exact items: Next, ask for a list of your top 10 most and least profitable products. This gives you a clear "hit list" of items that need immediate attention.
Prompt: Show me the top 10 most profitable and least profitable products along with the percentage of profit and loss.



Step 2: Figure Out Why They're Losing Money
Now that you know what products are problems, the next question is why. The most common reason is that discounts are too high. Here’s how you can prove it with data.
- Connect discounts to losses: Ask Terno to plot the average discount against the profit for all your products. You'll likely see a clear pattern: the products losing the most money are also the ones with the biggest discounts.
Prompt: For each Sub-Category, plot their average discount on one axis and their total profit on another. Highlight the sub-categories with negative profit.


– What it is: A single scatter plot showing total profit vs average discount.
– What it shows: It compares the Average Discount for each sub-category (on the horizontal axis) against its Total Profit (on the vertical axis). Dots are colored based on whether they have a "Positive Profit" or "Negative Profit".
– Purpose: To prove the link between discounts and losses. The reader can clearly see that all the red "Negative Profit" dots are clustered on the right side of the chart, indicating they have high average discounts (e.g., 20% or more).
- Find your "tipping point": How much discount is too much? A simple chart can show you the exact point where your profit disappears. For example, you might find that any discount over 20% means you're losing money on that sale.
Prompt: Plot profit margins across different discount levels.


– What it is: A bar chart showing average profit margin vs discount.
– What it shows: This chart displays the average profit margin (in percent) that the business makes at different, specific discount levels (e.g., 0%, 10%, 20%, 30%).
– Purpose: To find the exact "tipping point" where discounts become unprofitable. The chart makes it obvious: any discount of 30% (0.3) or higher results in a negative average profit margin. This gives the reader a clear, data-backed rule.
Step 3: See How the Problem Affects Your Whole Business
Bad discounts don't just affect one product; they can hurt your business in other ways, too.
- Check your map: Are certain sales regions less profitable than others? A profit map can show you that some states with high sales are actually losing money, likely because of local discount strategies.
Prompt: Visualize a map of sales and profit across different states.


– What it is: A pair of geographic maps of the United States, shaded like heatmaps.
– What it shows: The map on the left, "Total Sales by State", shows which states have the highest sales volume. The map on the right, "Total Profit by State", shows which states are the most (and least) profitable.
– Purpose: To see how the discount problem affects the business geographically. By comparing the two maps, a reader can spot states that look good on the sales map but are actually losing money on the profit map, likely due to aggressive local discounts.
- Know your customers: Who is buying these heavily discounted items? An analysis might show that these deals attract one-time bargain hunters instead of the loyal, high-value customers you want.
Prompt: Do high-value customers buy different categories compared to low-value customers?


– What it is: A grouped bar chart.
– What it shows: It compares the "Sales Share (%)" of "High-Value" customers (red bars) versus "Low-Value" customers (blue bars) across the three product categories.
– Purpose: To show who is buying what by contrasting your two main customer segments. The plot reveals that "High-Value" (loyal, high-spend) customers buy significantly more Technology , while "Low-Value" (discount-seeking) customers buy more.
Step 4: Create a Smarter, Data-Driven Discount Plan
Once you have the facts, you can build a plan to fix the problem. Terno can help you create clear rules for discounting.
- Make smart decisions: Get specific recommendations on which products should get discounts and which should not. This allows you to run promotions confidently, knowing you're not losing money.
Prompt: Recommend which products should get discounts and which should not.



Conclusion: Price with Confidence, Not Guesswork
You don't have to guess if your pricing and discount strategy is working. By following these steps, you can use data to find exactly where you're leaking profit and build a smarter plan for growth. It's about turning your business data into your most valuable asset.
Prompt: Summarise the entire findings and visualise the conclusion.



– What it is: A final dashboard that combines three mini-charts to summarize the entire analysis.
– What it shows:
- Profit Margin by Region: A bar chart showing that the West and East regions are most profitable, while the Central region is the least.
- Category Sales Share & Margin: A combo chart showing that Technology has both high sales and high margins, while Office Supplies has high sales but a low margin.
- Customer Segment Counts: A bar chart that shows the number of customers in each segment, highlighting a large group of "Discount-Seekers".
– Purpose: To tie all the findings together in one place. It gives the reader a concise, visual guide to the main strategic recommendations: focus on high-margin regions (West, East), invest in high-performing categories (Technology), and cultivate loyal customers over discount-seekers.
Useful Links
- Terno AI: https://terno.ai
- Dataset link: Superstore Dataset
- Detailed analysis: https://nikita.app.terno.ai/chat/share/0aa37aca-52d6-4bc2-b7e5-c1e45963e4f6