How we created an internal Spotify Wrapped analogue in less than 2 weeks and almost 0 code.

This is the story of how a team of 3 built an end-of-year summary for a product studio of 200+ people using CSVs and low-code tools. The result looked like this

Maybe it could even be productized, as end-of-year summaries have been quite popular last year.

I will write about:

1. The concept & how we used Spotify's generative approach to build our campaign (and you can too!)
2. Everything data — how, where, when
3. The final result

Here we go!

The concept & generative approach:

Our concept was simple — we wanted to have 2 things:

1. General data on everyone combined — through a landing page
2. Personalized data for each person — sent through email and our Slack bot

For the landing page we’ve used Readymag, we just needed to display the data in a nice way.

For the personalized data, we went with Sendgrid for the emails, as we have already implemented that for one of our internal products, as well as Slack bot that we already had within that very product.

Now we’ve figured out the concept, now time to figure out how personalized data should look like.

Here comes Spotify’s generative approach:

Essentially, each card for the person is a random mix of:

1. Geometric figures
2. Background
3. Colors for the two above-mentioned elements

You can simply mix and create a lot of unique cards. That’s neat, right?

Here’s the base we’ve come up for our summary:

Now all that remained was to get the data for both the landing page and the personal cards!

Everything data — how, where, when

We’ve used 4 data sources:

1. Slack
2. Harvest - time tracking application
3. Everbase - PTO/HR software
4. ParaMerch - our internal eCommerce merch store

Luckily the data all was exportable, the only difficult thing is to glue everything together, considering Harvest data had 98k lines.

For that very task we’ve used Retool Workflows and it looked like this:

Looks scary, but in reality, it’s pretty simple:

1. We extract the data from exported Google Spreadsheets tables
2. We filter the data to adhere to what data we actually want to get
3. We aggregate the data for each table by project/person in our case
4. We get all the data from each separate table
5. We create a new spreadsheet for the new data
6. We populate the new spreadsheet with the data collected

It worked wonders, considering the Harvest spreadsheet was around 98k lines of data!

What was left is to connect backgrounds, geometric figures, and colors from the previous paragraph with the data from this paragraph!

The end result

Here’s how the personalized card turned out:

And that’s how we’ve built the end-of-year summary for our product studio!