DEV Community: Ruggero Motta

Prices Dataviz: Auchan vs Monoprix

Ruggero Motta — Thu, 28 May 2020 17:33:23 +0000

A data analysis through data visualization

This is a price analysis of two main supermarket chains operating in France.

I got the prices online, imported the data in R for data analysis and plotting. I then exported the most significant visualizations to show the insights.

Why I did this

Of course, I wanted to compare the offering of the two grocery stores to be able to do more informed choices, but I also did it for curiosity.

Specifically, I was curious about the relation between their brand positioning and their pricing strategy. In other words, I wanted to discover if my perception of their brand was consistent with the prices they offer.

Least but not last, I love information visualization, I wanted to play my R and data-analysis skills, and to try out the new Affinity Designer app.

Hypothesis

My perception was that Monoprix has a generally higher brand positioning than Auchan. This was because of the retail experiences as Monoprix comes with cleaner and brighter spaces, but also because it provides a more consistent and original visual identity in all of its customer’s touch points.

Having tried both, I also had the perception that the groceries were a bit cheaper at Auchan, but I never had the chances to compare the same products: they often sell different brands, and commodities brands like CocaCola have the same price everywhere.

Getting some Data

The main concept here was to get the prices for several products from different categories. To be able to compare the stores I needed a clean list of homogeneous products with their own prices.

I’ve run some manual queries over the websites and I soon realize that to avoid spending too much time in cleaning the dataset, It would have been better to narrow down my researches to highly specified categories. Indeed, searching for a product like Olives resulted in including pizzas with olives or other related products, but narrowing the search to Green Olives produced quite more clean results.

After some try I ended up with these trustable categories:

beers
eggs
green olives
UHT milk
minced meat
tuna (in water)
spaghetti

I got the data from the Auchan and Monoprix online grocery store and in just one hour I had my dataset ready to be plotted in R.

The number of products I got for each category in each store vary between 12 and 35 items. The beer makes exception, with more than 150 records from both the stores.

The main property I targeted was the price per Kg, which becomes price per liter and price per piece for the milk and the eggs respectively. To make the items comparable it was necessary to consider this prices instead of the canonical items prices. Luckily both the stores provided this information.

Data analysis

The first interesting insight was the different offering strategy the two supermarket use for the beers. Both of them have a huge offering in terms of items, but the way these items are distributed over the price axis is pretty different.

While the majority of Auchan’s products are priced under 5 euro and its whole histogram results shorter and more compact, the Monoprix one is longer, with a more balanced presence both before and after the 5 euro line.

The highest prices of the category are around 11 and 12 euro per liter and are provided by Monoprix. On the other hand Auchan seems to cover better the bottom end, with more items between 1 - 2 euro per liter.

The insight seems to be confirmed by the other categories.

In all of them, Monoprix tends to cover a wider range of options along the price axis, included the most priced items in the category, while Auchan’s offering is more dense in the low-cost area.

Further confirmation were observable from a facet plot like the one below, where we can appreciate the position of every single product over the price axis.

Prices of commodities products, like eggs or milk, are clearly compressed in a small area at the bottom of the plot, indicating that for this kind of products it’s really hard for the store to operate a real diversification. But as already pointed out, also here Monoprix is still providing a more variegated offering and Auchan looks to be focused in providing the better price.

Again, is important to notice that the number of items is not relevant here and the fact that a store has more (or less) items in a category is due to the way I created the dataset. It doesn’t mean that a store offer less products than the other.

What’s important for this analysis is that every store has enough products in each category to give us a trustable representation of the stores’ offering in terms of price diversification.

Indeed, to have a cleaner view of the situation, we can use a box plot, which is based on the quartiles of the series and help us to not taking in account the number of items taken from each store.

Now we can really appreciate the different strategy the two stores are putting in place. The central quartiles of every category show us Monoprix with a definitely higher priced offering and Auchan with a quite more compressed offering, always closer to the bottom.

Even in the eggs' category is now clear how the Auchan offering, by a pure price perspective, is more competitive than Monoprix.

Conclusions

My original hypothesis about Monoprix was confirmed: my perception of a higher brand positioning is reflected in higher priced products.
Auchan is more competitive on lower prices, having a quite bigger offering at the bottom end of the price axis, while Monoprix provides a more variegated offering (from a price perspective) which includes also some more expensive products. Moreover, in the majority of the categories, Auchan provides the absolute cheapest options.

Is worthy of note that, how we can see in the histograms, almost no category tends to follow the normal distribution. Instead, each category across the two stores, seems to follow a specific distribution curve.

Thank you for reading!

I’m a designer with a passion for data science, but I’m not a data scientist. If you are, and you notice any mistakes in the terminology or in the analysis, please add a comment below, I’d be pleasured to read constructive criticism or to discuss alternatives approaches.

Notes

The R plots were imported in Affinity Designer for graphic editing.
Data sources: https://www.monoprix.fr/courses-en-ligne https://www.auchandrive.fr/

Designers and E2E tests: Dumb bots are our friends.

Ruggero Motta — Mon, 27 Apr 2020 07:39:39 +0000

As designer, we know that user tests are crucial for creating a successful application, a product that covers the users needs, keeps the user frictions as low as possible and, for the B2C market, increases the conversion rate. Keeping in mind that this kind of tests are incredibly important for a product designer, we should also be aware they are not the only ones that a digital product requires.

Indeed, when we design within a continuous delivery environment, we should consider to not testing only the finished product but also taking care of the intermediate versions built sprint after sprint. Each of this version is a step on the path toward the next release and if we want to make this path smoother and safe(and if we care enough about our designs) we should run tests on it.

As I recently had the chances to work with the Mikamai's and the Wizkey's team and to gain some experience in testing the software, I'd like to share here my insights.

This is an article written from a product designer perspective and I will not write about technical approaches or tools. Instead, I'll write about how E2E test can help a designer creating better products and increase his/her contribution within the agile team.

Testing is the new religion.

Modern software development thinks very highly of tests, and indeed the Agile methodologies impose several tests. Many of them are "automated" which means they are automatically run by a script. They don't need a human being clicking around the screen, instead they need a human being writing the script.

The most interesting for a Designer is the end-to-end test, where the two 'ends' are the user behavior and the expected result, respectively. In synthesis, we perform a user flow on the app, targeting a specific feature, and we check whether the result is the one expected. For example, if we click the "logout" button we expect the session terminates with a logout page loaded, if this doesn't happen, the test fails.

Pretty simple, no? It's actually so simple that it can be automated.

"Automation is cost cutting by tightening the corners and not cutting them."

-Haresh Sippy-

What is an automated end-to-end test.

Imagine a dumb bot programmed to open the application and to interact with it, performing a specific user flow at an incredible speed. Then, it checks if the result is the expected one and more specifically it checks if the UI shows what it's expected to show. If it doesn't, it tells us where it found the error.

The video below is a screencast of an E2E test suite we ran at Wizkey. It's made of various tests that cover some user flows, starting from the shortest ones (e.g. logging in). When a test is successfully completed the console shows a green tick with the name of the test.

Considering the bot takes few minutes to test all the main product features and that it can be launched at any moment on every version of the product, it becomes clear how useful it is.

A designer's job.

The main reason of the fast growing adoption of E2E tests is the wide range of effective tools available today. In the recent past, tools like Cypress, Test Cafè, Puppeteer and many others, made the E2E testing more accessible and less expensive. The game changer is that some tools now provide a clickable user interface with which we can create an automated test even with very basic coding skills. This opens a door to all the not-devs people that work in the team, like product owners and of course, designers.

But why we should get our hands dirty with this tech sauce?

Every team has a different organization but an undeniable truth is that the designers know more the user flows than the others, because they are the ones who designed all of them. As they are often close to the business team, they have a wide and forward-looking view on the product. They can better understand which user flows are the most important and what are the key points they should not do compromises on.

In a scenario where the product is being built from scratch, we can use E2E test to advocate user needs: we can write a test to highlight a specific Acceptance Criteria that we consider essential for the usability, to make sure it will be implemented in a certain way. Hopefully, this will make it a shared interest of the team and it will not end up being half forgotten within the sprint hurry.

"Any fool can know. The point is to understand."

― Albert Einstein

The most interesting part is that to write a test we probably need to have the development environment installed on our computer: to create a test, run it and then push it to the git repository. The only fact to contribute to the code base will help us tremendously in understanding the product mechanics. In fact, we can learn a lot by talking with a developer about a failed test: we can discover technology constraints, opportunities, and even come out with new solutions.

Moreover, the first reason why I suggest every designer to have a look to the E2E tests, is because by creating a common ground with developers, we'll improve our communication. All this will make us a better and more incisive designers. I'm not talking on learning to code and be able to pair programming with a developer, I'm talking on building bridges.

Writing a test and using a git client to push our 3-lines-commit to the project repository is also a try in creating a common ground.

Conclusions

Digital Product Design is a very multidisciplinary field and more and more specialized roles are emerging. We can't master all the aspects of it, but we can develop several side skills that can help us in doing some good design and to communicate well with our team.

For sure, learning to write an e2e test tremendously increases our ability to understand the product we are designing, to find *feasible solutions and it creates a common ground within the agile team.

"Study hard what interests you the most in the most undisciplined, irreverent and original manner possible."

― Richard Feynmann

Ipad pro's box, recycled! 🤯

Ruggero Motta — Thu, 23 Apr 2020 22:15:03 +0000

My girlfriend ordered the iPad Pro and it was finally delivered yesterday.
The cardboard box had "Please Recycle" written on it.

It took around 10 minutes to build this awesomeness:

Thanks to the add-on it comes in 2 different heights 🤩
The very cool part is that since the lockdown we really needed a second laptop deck.
It was just a pity she had to use hot glue to join 2 pieces, glue is never environment friendly ... but as we are under lock down we have to do our best with what we have in the house.

PS: I'd like to share the instructions but I didn't see the building process.

Have you ever checked the Double Fine website?

Ruggero Motta — Mon, 20 Apr 2020 22:16:29 +0000

I just landed on the Double Fine website. It's amazing, they don't give a F. 😂

Is a Data Science bootcamp a good option for a Designer?

Ruggero Motta — Sun, 19 Apr 2020 21:23:07 +0000

Ok.
I'm a ux designer,
and I'm seriously considering a Data Science Bootcamp.
I have some technical background. But I'm not sure this would be enough to successfully complete the bootcamp.

Did you guys ever tried a Data Science bootcamp?
If yes, Could you please provide me some insights on that?

Totoro is magic

Ruggero Motta — Wed, 15 Apr 2020 22:05:39 +0000

Inspired by Myiazaki's Totoro in dusting off my CSS skills, I ended up with this interactive animation in pure CSS.
I've learned that is actually possible to manage states also in CSS, through the :target property, without a single line of js.

It was a very funny exercise for me, so I propose a #CartoonCSSchallenge:

Pick a frame from your favorite anime or cartoon and enrich it with some good CSS sauce 🍄, create some magic! 🧙‍♂️🦄