20 Years of Sales Data

Hi all,

New CEO just asked me for 20 years of sales data monthly by brand & product.

This is literally hundreds of products and about 40 different brands over the span of 20 years.

I’m looking for some ideas on how best to display this data, first at a summary level and then being able to drill into detail if need be.

I know my request if pretty general but I’m hoping someone has done a similar thing like this before and has some good guides or charts that I can draw inspiration from?

Many thanks,
Tim

1 Like

Hello Timmay,

From your post I understand that the volumetry of your data will be around 3 million rows
20 years of data x 12 (by month) x 40 (brands) x 300 (hundreds of products)
The volumetry is important to consider, because based on the connexion type import/direct query/live connexion and your model, the visuals can be more or less fast to load and can influence the way to design your report

Before jumping into the visualization part, I recommend you think about the audience, build your data story and decide the user journey.
You said that the report is for the CEO. What does he/she want to know? Should it be an exploratory report, does he/she have an idea of the questions to be answered? For ex, for this type of data, if I were a CEO I would like to see the evolution in time, compare the products to a previous period, compare products between them, look to see if it is a seasonality. I would also want to see the trend, are sales going down, should be products be replaced or not? As a CEO I would also try to go to the essential. Do we actually need 40 brands and hundreds of products?
Once you have the ideas and questions to be answered you will need to perform a data analysis/discovery, think and sketch.
You create a report for the CEO, your design needs to be clear and straight to the point. Do not include 100 different findings, concentrate on the essential ones.
So, my recommendation is to explore your data, find all the findings that can be interesting and at the end choose only the ones that answer the best the requirement. Think about the user journey. Think about navigation, buttons, interactions, tooltips.
Do not use sophisticated charts. Use bar charts, matrix, cards, or any chart that can clearly express the message you want to send. Be sure to add context with clear titles and subtitles.

If you are looking for inspiration, I recommend looking at Power Bi Challenge N.10. The requirement is somehow close to what you need to create. Look at the reports, read the write ups, I am sure it can help you.

Good luck,
Best regards,
Alex Badiu

5 Likes

Hi Alex,

He said the main purpose for seeing sales history is so he can build ideas about forecasting.

Thanks for your insightful questions, I will check out challenge 10!

Thanks,
Tim

1 Like

Hi @Timmay, did the response provided by @alexbadiu help you solve your query? If not, how far did you get and what kind of help you need further? If yes, kindly mark as solution the answer that solved your query. Thanks!

2 Likes

The challenge he referred me to wasn’t related to my question so his answer didn’t really provide much insight. Is there any material on Enterprise DNA that can assist me at all?

Did you do a search within the forum? Did you do a search on youtube? Did you do a search on the internet? Did you do a search on Microsoft Power Bi community?

@Timmay,

I recommend looking at the challenge n10 for the following reasons:

  1. You said you have 40 brands and hundreds of products. In challenge n 10 we had to work with a lot of data combining materials/vendors, vendors/plants based on historical performance.
  2. I thought you can find inspiration. You said in your request you are looking for ideas: top 10, worst 10, anomaly detection, forecast, navigation, buttons, color themes, data story, scatter charts. All these you can find in the reports that were posted. And you have the write ups too. And you can download the .pbix.

In my opinion, you can find all the ideas and knowledge you need to build a great report.

Best regards,
Alexandru Badiu

2 Likes

Hey @alexbadiu I had a similar report/issue where I had to display lots of KPI’s/categories and I used Challenge 10 for inspiration. Used Top 10’s, Worst 10. Also found the Decomposition Tree useful to give the user some power to how they viewed the data.

The challenges are a great source of inspiration for reporting challenges.

1 Like

Hi Timmay,

I’ve experienced similar requests at previous positions. Generally a request for a data dump, as much info as possible across a specified time frame and the requestor has an idea that they can look at all of this granular data to discern a pattern or some sort of relationship.

You mentioned that your CEO said his main purpose is to see sales history so that he can build ideas about forecasting.

What I would do as an analyst if I received this request would entail two parts: 1) the detail he asked for and 2) some highlights summarizing any analysis I performed.

First, I would pull the initial dump and then maybe make a table for the brands and products, perhaps a hierarchy - with summary stats for the current FY and some kind of indicator to show if current stats are in line with history or have increased/decreased.

This hierarchical table could be used as a selector for some line or bar charts to show the trend over time. You could even add filter on the page for the ‘look back’ period - with selectors for last 5, 10, 15 and 20 years.

I might also experiment with the forecasting tool that is included with PowerBI. I think it only works with line charts that have a single series. If your CEO is anything like the CEO at my last organization, he likes “snazzy” things and he did say that he was building ideas about forecasting.

So that is part 1 - it’s pretty much what was asked for with a layout that makes it easier to look at trends over time by brand or product. Maybe even the ability to sort by those measures (sales, units sold… whatever is important to your business - I’m a healthcare analyst so I’m not really sure what people in the sales field are interested in - but you’ll know what your org measures :slight_smile: ).

Part 2 is the the analysis. What I would do is using this charting tool and the summary statistics as a starting point, I would look into the products that have changed to see what’s going on with them. My goal would be to provide the CEO with a higher level summary analysis. Very importantly I want to provide him with statements here - not more questions.

I would want to highlight products or brands with actual changes vs changes that might have been driven by black swan events such as “product X sales have been down compared to historical norms, but that coincided with a drop in production tied to supply chain issues as a result of that tanker stuck in the Suez Canal” or “toilet paper sales increased in FY2020 compared to that last 5 years, but that was primarily driven by panic hoarding in Q2 during the pandemic”.

I would also want to pull out the products or brands that have had changes, compared to (what I would assume would be) the majority that’s just trucking along, business as usual. I would want to filter out the noise of business as usual and special events in able to focus on brands or products with indicators of true change.

These pull-outs/highlighted products might each have a different set of presentation charts. Which you’ll probably be able to select from because you’ve had to use more of these in your analysis.

As others in this thread have mentioned you need to be brief. Hit 'em with the highlights, as one of my managers used to say. And remember, BLUF - Bottom Line Up Front.

You can them give him a super-bullet point summary along with the data dump he requested, but hopefully that data dump is set up in a way that lets him filter and look at details he might be interested in via the hierarchical table with summary stats and the linked trend charts.

FWIW, I’ve been through similar scenarios with new execs numerous times - either because they are new or when I’ve been new to the analyst team. Most of the time I’ve been able to develop a very positive working relationship with them where the exec starts actually talking with me and letting me know more of their ideas and the questions they want to answer.

Data consumers generally haven’t learned how to articulate their questions, and are used to just asking for a data dump. I’ve found that giving them what they’ve asked for (which is always important, I’ve known people to get frustrated when they don’t get what they asked for) as well as some sprinkles of insight work well to move toward a more collaborative analytical partnership.

Sorry this answer was probably a data dump in itself. I hope it’s helpful.

5 Likes

Thanks Karen, that’s very helpful.

1 Like