How we built our 'What Musician Are You?' Instagram filter

Written by Benjamin Davis

This filter started off as a fun little project but it ended up being quite an interesting venture and the process to make it used a few cool tools that I wanted to talk about. One of which is our own Rankd API that I have been building in the background for a while now and the other is a tool called Robolly. These tools allowed us to create a filter that included images of over 150 musicians.

But first…

Why create an Instagram filter in the first place?

I’ve made quite a few features for my personal Instagram and marketing clients of mine. I’ve seen first-hand how they can be great at driving engagement and lots of free awareness and traffic to an Instagram page.

Instagram filters are great to promote brands, increase awareness and increase engagement with a following. We thought this was perfect as we could grow awareness for Rankd as well as awareness of some local musicians along the way. It could also show off some cool features of the Rankd tool which I’m discussing in this article.

We wanted to experiment with this and Rankd. Our idea to create a ‘What are you’ filter was partly because I had the background setup of the filter already and partly because we could show off some features of our database.

How did we create so many images?

We wanted our filter to scroll through images of all, or failing that, as many musicians as we could. But creating this many images would be challenging.

Creating 166 images could be very time-consuming if we were making these all manually so I wanted to find a better way. As we already have the names and profile pictures in our database, I thought we could combine this information with a graphic template. This is where Robolly comes in.

What is Robolly?

Robolly is an image generation tool that can bulk-make images from a CSV upload. To get this to work in our scenario we just needed the image template:

And then a list of musician names and profile picture URLs:

Using these we were able to create our 166 images. We then just needed to optimise them to fit within the 4 MB limit for Instagram filters. For this I used TinyPNG which is my favourite website for quickly compressing a lot of images.

How did the filter go?

Great. Performance in the first few days was far beyond what we expected.

In the first 3 days, the filter had been opened by 4,300 people, used 2,800 times and been seen by 9,500. The number of local bands we saw using and sharing our filter was amazing. It also sparked a lot of conversations with musicians which was really nice to see.

This filter also helped us have the best month yet in terms of social performance.

Issues we encountered along the way

Unfortunately, after launch, we ran into some issues with a planned update that would add many more artists to the filter. For some reason, Meta thought that our filter was promoting alcohol, tobacco, weed or other drugs.

We kept appealing this but the issue came up a few times for some reason.