Previous
¶ Analytics
¶ What it is
Analytics is a data source that gives you information on the existing behaviour of customers across touchpoints to that you have access. Looking at Analytics, such as click paths, bounce rates, screen heat maps, and similar, helps you understand how customers interact with existing touchpoints on a large-scale, quantitive basis when they do not feel like they are being observed.
¶ Why it is useful
Analytics can give you insight into how customers behave today, what they are interested in, what they pay attention to, and where they end up spending time and clicking. It does not tell you anything about their motivation to perform any of those actions or evaluate alternative scenarios. It means you know what is happening today.
¶ When to use it
Suppose there are existing touchpoints you have that are or will be related to the product or service offering you are looking to create. In that case, you should always have a look at the Analytics of those to understand existing behaviour.
¶ How is it done
- Understand which existing touchpoints are or will be related to what you are looking to offer customers and where target customers are already interacting.
- Request the Analytics for those touchpoints from their owners, explaining the scope and scale of your inquiry and the reason for it.
- Sift through the data in different time horizons and work with a data-literate analyst to slice and dice the data to get various types of information and statements about what users are doing today.
- Start diving deeper, e.g., looking at demographic segmentation, drill downs, and correlation analyses to get at further statements about present behaviour that can be useful.
- Discuss relevant insights and conclusions with the team.
¶ Do's & Don't
Do's
- Try to get datasets with relevant scale and scope regarding rows of data and duration of data collection.
- Keep a master copy of the original data before running analyses and transformations on the data.
- Do the common sense checks or sanity checks to reflect on whether a conclusion makes sense and at what level it makes sense.
Don't
- Avoid drowning in data by asking for way more than needed (e.g., ten years of data if the touchpoints were redone a year ago and the previous nine years won’t tell you anything).
- Don’t confuse correlation with causation - it is rarely clear what causes what, which might be a question for qualitative customer interaction.
¶ Tools needed
- Access to relevant analytics accounts
- Excel or Google Sheets
- Statistics Tools (R or similar)
¶ GenAI Tool Use Case
¶ Quickly Translate Your Data into Descriptive Text
Users can swiftly create report pages by providing instructions to PowerBI Copilot, enabling them to efficiently transform data and charts into textual descriptions and summaries.
Key Steps Tutorial:
- Create a report page
- Choose the dataset you want to explore and click “More options” > “Create report” > Click “Copilot”
- If you already have a clear plan for the content of the report, you can simply click on "Create a report that shows…”; if not, then click on "Suggest content for this report”
- After you have revised and refined your requirements, click “+ Create”
- Sales performance by product: "Create a page to analyze the sales amount, revenue, and profit margin of different products, categories, and subcategories over time and across regions."
- Generate data descriptions and summaries
- Select “Narrative” in the “Visualizations” pane
- “Choose a narrative type” > select “Copilot”
- In the "Create a narrative with Copilot" dialogue box, choose one of the options, or edit your own prompt and select the range of reference visuals > click “create”
- Adjust the narrative by changing the instructions for Copilot, or using the suggested prompts
- “Shorten this summary and bold the key information”
- "Make the first bullet point about length of stay."
- "Generate a summary explaining the relationship between revenue, location, and primary interest.”
Sample Prompt: (Provide by official)
Sample Prompt: (Provide by official)