Discover Relevant Insights Into Complex Problems
What Happened to Our Margins?
Your company’s margins are shrinking and you're not sure why. But with Einstein Discovery in your toolbox, it's time to create a story that will give you answers!
Create the Story
- From the dataset's menu, select Create Story. Analytics Studio launches the Story Setup wizard.
- Based on the data in your dataset, Einstein Discovery suggests creating a story to “Maximize the field Margin.” The story is designed to answer a question: How do I maximize margins? And that question lines up with the most pressing question you have: What happened to lower our margins? You can open the lists below I want to and the field to see the other available selections. But in this case, Einstein Discovery is exactly right, so don’t make any changes. The default story name is the same as the dataset name, but “APDist” isn't descriptive enough. Change it to “Maximize Margins”. Keep the default App as “My Private App”.
- Click Story Type. For our investigation, we want Einstein to conduct the more comprehensive analysis of the data. Therefore, click Insights & Predictions.
- Click Setup Options. For our investigation, let's have Einstein Discovery select the relevant fields for us.
- Accept the default (Automated) and click Create Story.
Now sit back and watch as Einstein Discovery evaluates your data from just about every possible angle. This analysis usually takes a minute or two, so go ahead and take a stretch break.
When it’s done preparing the story, Einstein Discovery shows you the insights it found.
Here are the key areas of the interface for Einstein Discovery stories:
|Story Headline||Name of this story, selected goal, most recent version.|
|Story Toolbar||Tools you can use to view predictions and improvements, update story settings, and other tasks.|
|Variables Panel||Shows you the list of variables in your story and their correlation to the story outcome.|
|Story Version Summary||Summary of story insights, including version comparison.|
|Insights List||List of insights associated with this story.|
Initially, you see descriptive insights in the insights list. These insights show what occurred in the past. The first visualization you see—a bar chart with highlights and explanatory text—represents the first of many insights that Einstein found. In fact, it’s the most statistically significant insight in the list. You can scroll through the list to explore other insights. Different products have widely different margins. An Einstein Discovery insight doesn’t simply present some data, it also highlights what’s important about it. The blue bars show the products where the margins were significantly better or worse than the average margin (the orange horizontal line).
But that’s a discovery you could have made easily enough on your own. The question is, why? What’s the cause? There are more insights in the story — let’s focus on the product with the lowest margin, alternators.
- Scroll down to the insight with the title “When Distributor is Nisizu, Product: Hybrid Motors and Lift Supports do better”.
Great—looks like some products are returning good margins, particularly Hybrid Motors and Lift Supports, as the insight shows us. But it still doesn’t tell us why alternator margins are lagging so far behind, so let's keep going.
- Click Product - Alternators listed under the header “Cases where Nisizu did worse than others:”. The new insight tells you that all distributors are under-performing with alternators, but there’s one that’s doing especially poorly: Nisizu. You suspect that this distributor’s poor performance has had a major impact on overall margins. Let’s take this one step further and find another insight that can tell us why.
- Above the insight, click Why does this do worse? This kind of insight helps you take a deeper look into the exact factors that led to a particular outcome.
You see a new insight that supports what you already suspected: “Product is Alternators by itself explains 4.123 of the change in Margin.”
The waterfall graph reveals a new view of your data. The explanatory text says: “Distributor is Nisizu occurs 35.8% of the time globally but it changes to 52.8% when it is known that Product is Alternators.” In other words, Nisizu is responsible for more than 50% of your distribution deals for alternators! That explains a significant part of the drop in margins.
Next Question: How Can We Improve Margins?
Now you have some indications about what’s causing margins to fall. That’s great, but you need a solution, so the obvious question to ask next is “What can I do about it?”
- Predictions of future outcomes based on a predictive analysis of your data
- Suggested ways in which to improve your predicted outcome based on a prescriptive analysis
Let's take a look.
- On the Insights navigation bar, click Predictions.
- In the left panel, click Product and select Alternators from the list. The projected margin is 20.68.
- From Distributor, select Nisizu. The projected margin decreases to 15.02, which reaffirms what we already know - that distributing alternators through Nisizu lowers our margin.
- Select different distributors until you find one that provides the highest margin for alternators. Consider whether to redirect alternator sales through this distributor.
- Let’s see how Nisuzu performs with a different product. For Distributor, select Nisizu again.
- For Product, select Belt Drives. The predicted margin increases to 56.90. Wow! That’s quite a jump! However, before we conclude that distributing belt drives through Nisizu is the best option, let’s try something. To see what the margin for belt drives looks like across all distributors, remove the Distributor filter.
- For Distributor, click Select an Option (which clears the filter). The margin drops to 49.48. Having Nisizu distribute belt drives improves margin.
- Select different distributors until you determine which one provides the highest margin for belt drives. Perhaps you can consider redirecting belt drive sales through this distributor.
- Salesforce Help: Explain, Predict, and Take Action with Einstein Discovery
- Trailhead: Einstein Discovery Stories
- Trailhead: Einstein Discovery Story Insights