I Tried to Automate a Business Dashboard — It Didn’t Go as Planned

This is not a story of success.

This is not a story of success.

I wanted it to be. I wanted to tell you how to do this. But it didn’t go as planned.

Last December, I set myself a holiday project: I really wanted was a statistic dashboard.

Really ambitious of me… But well — I’m stubborn and in possession of a “how hard can it really be” mindset. So I started to craft my plan anyway.

Phase 1 — What do I really want ?

Phase 1 and 2 are my favorites. I always enjoy more the planning and research than I do the building, or even the result. It felt like a Christmas present to myself — to start scribbling on my iPad. And to imagine what I was going to build.

So I wanted a statistic dashboard.

Something that would show me how many sales I made this month, and the last months, my impressions and views on social media and how many pieces of content I had created.

The goal was to motivate myself to push harder, to create more, and to have a good overview of my financial and content performances.

I also wanted to be able to see this on all of my devices. Including my phone.

I had in mind something like this :

Now…

Phase 2 — Is this even possible ? And how?

Dreaming is easy. And I know automating looks like an El Dorado of opportunities — but is what I want actually possible?

I would need to retrieve 3 kinds of data :

• Sales data (who, what, when, where, how much)
• Social data (impressions, like, views, clicks, followers)
• Creation (number, type, destination)

And then compile that data into conclusions.

My first question was : where is that data?

I figured the best way to retrieve sales data was from the payment processors : so Stripe and PayPal.

Creation data was in Notion — or, directly on the social media sites: Twitter, Medium, YouTube, Anchor (my podcast host) and MailerLite (my emailing service).

If that seems like a lot, it’s because I run several projects online. My online school of French, my French-speaking technical coaching brand and my Circé Creates project.

Social media data was also in the destination sites.

Now, where should I be taking all of this data to?

The most flexible place I could think about was Google Sheet. I could always send that data elsewhere after.

Final question, who would be the link between Google Sheet and all those sources ?

After a bit of research, I landed on Zapier. I had already used it a bit, and as far as I could see, all my sources were in their catalog of apps.

Here was my plan :

I was ready to move on to phase 3.

Phase 3 — Start building (while expecting set-backs)

This is the phase where you have to be strong, and where something is going to get messier than you had planned. This project delivered hard on that last part….

I decided to start with the sale data.

Because there were only two sources. So in theory, it would be simpler.

Setting up my automations was a tiny bit harder than I thought. I wasn’t expecting this much mumbo jumbo. And I had to make a few adjustments in the Google Sheet itself.

Especially because I take payments in both euros and US dollar. And so I needed to somehow combine the two devices in a way that made sense.

I couldn’t find a way to automate selecting the right project (School of French, Circe creates…). But that wasn’t much to do manually at the end of the month. Eventually, I managed.

I tested it all — and yes, it seemed to work fine. So I went to sleep. Waiting for the next sale to see if it worked fine for real. When I woke up, it had worked as expected. Marvelous!

I had my sales data automated in Google Sheet.

Second set of data : creation data.

The setting up was faster, and easier. Except for YouTube — whose API apparently was acting out. No matter, I don’t post much on YouTube.

I had most of my content automated in a neat list in Google Sheet.

I went to sleep again. Dreaming about how I could rule the world with such a power at my fingertips.

I woke up to… bugs.

Phase 4 — why is it so damn complicated ?

My automations had stopped working. Both of them. And the troubleshooting messages were about as useful as soft rubber teeth.

No worries, I thought. Google knows it all.

So I started digging. And digging. Until I was up to my neck in territories I didn’t even understand the name of. It was all so very technical.

I worked on my keywords, looked on YouTube, Reddit, Twitter and even Quora.

And let me tell you, when you go this deep to look for a solution, there comes a point where you have to decide if you are just an idiot or if the rest of the world is conspiring against you. I usually decide I’m an idiot.

Idiots make a lot of basic mistakes. And so I checked my basics again and rebuilt the automations from scratch.

It worked.

But I still had no idea why it stopped in the first place. And I certainly had no confidence this was going to keep working for more than a couple of days. I kept going anyway.

I still had one type of data to source : social media performances.

Which proved to be much harder to find than the other two types. Sure, with Zapier I could post videos on YouTube, and compile lists of tweets — but there was nothing about performances. I wanted the number of views, the replies / comments… But I couldn’t find anything.

I went to sleep thinking about how ruling the world was a very hard job.

I woke up to my automations half-working.

Off with my own head!

Phase 5 — Is this even worth it?

« How hard can it be » is a good mindset to have. But sometimes it crushes my soul a little bit.

I started to reconsider my project. I could still do it manually. It would take a couple of hours over the course of a month, but I could log the sales and numbers myself.

Because, if I need to worry about my automations stopping all the time, is it worth automating? I decided that, no.

Is there an app that already exists?

Turns out there are a lot of KPIs dashboard solutions. But I wasn’t willing to pay what they asked for. Or they weren’t a good fit for me.

So I don’t have an automated statistic dashboard.

This is NOT the end

Maybe I will find it in me to try changing different factors.

Try Make instead of Zapier. Replace Google Sheet. Or find another source to get the data from.

Maybe one of those will work better. Or maybe I will think of a complete different solution that could give me the same kind of result.

Like the Easel function from the Arc browser.

In the meantime, I am manually compiling my data. And I have learned a lot about automation and Google Sheet.

So this is not a story about failure either: