What I Learned Spending €100 on Online Ads

Yes, we're going to talk about online advertising. Don't run away :) I know it's not really the typical topic of this blog but it will be an excuse to talk about analytics, conversion, marketing, customer acquisition strategy and I promise, it's rather interesting.

Building a product, an application, a website, that's good. It's already not given to everyone but it's actually the simplest part. Because yes, the difficulty comes after, when it comes to getting known.

A good product that nobody uses is about as useful as an election platform the day after the election. Everyone couldn't care less.

(I hope this joke will please the person who criticized me for talking politics in my last blog post :))

Anyway, in this post we're going to talk about online advertising.

Paid acquisition

Paid acquisition consists of spending money to try to create traffic to your site.

Inside we can categorize several things (non-exhaustive list):

• SEA (Search engine advertising) which consists of placing ads on search engines via Google Ads or Bing Ads
• SMA (Social Media advertising) for ads on social networks (Linkedin Ads, Reddit Ads, Meta Ads etc..)
• Retargeting: targeted advertising, the thing that shows you diaper ads on all the websites you visit because you had the misfortune of looking at an ad somewhere once
• Sponsoring, which consists of paying for space in a newsletter, a video, a podcast etc… Yes, it's the famous VPN ads we see on all YouTuber channels for example.

I'm starting out, I have a small budget, so SEA and SMA are the two most affordable channels for me.

So I isolated Google Ads and Reddit Ads but not everything went as planned.

Conversion tracking

Most tools like Google Ads, Reddit Ads and the like will try to optimize the display of content to the audience that is supposed to "convert" best.

Yes, because, selling diapers to teenagers for example, that's not the most effective.

You're going to pay for displaying an ad but at any given time, many people also want to display an ad so you're in competition with them on a bidding mechanism.

By default you can just say "on this keyword, I want to bid a maximum of €3" and that's it.

But for several years now, most ads platforms offer you to put tracking mechanisms on your site to determine if the conversion really happened, if the user really made a sale, or a signup. And this signal will then be used by the platform to decide whether to bid or not.

In short, before we used to say: "I'll pay a maximum of €3 for this keyword" but now we give a maximum budget over a period of time and ask the platform to spend it and figure out on its own to make it as effective as possible. Platforms will analyze thousands of signals in real time to decide whether they bid €1, €5 or nothing at all because the user is unlikely to convert.

Except that, bad luck, on Writizzy.com I don't want to put any overly intrusive tracking tools that store private information and cross it on multiple sites. This is what allows me not to put cookie banners because I'm fed up with cookie banners everywhere and also because I don't want to base my product on selling my users' data.

Now I'm not saying I don't do any tracking, simply my objective is not to send private data to a third party and to calculate the conversion myself to adjust my budgets.

Conversion metrics

Here we're going to be interested in several metrics:

• raw traffic: the number of visits related to an ad campaign
• traffic conversion: are these visitors doing something useful for me
• customer acquisition cost (CAC), that is the ratio between the budget and the number of new users

For example if I spend 100 euros to get 1 customer, then my CAC is 100 euros.

It becomes interesting when CAC is less than LTV.

Ok, this sentence is cryptic.

The LTV (lifetime value) represents the value of a customer over their lifetime, basically the revenue generated by your customer.

For example, imagine a subscription at 9 euros per month. If the average duration of a user on your platform is 12 months, then the average LTV is 108 euros. In reality it's less, since you would need to subtract operating costs (the server for example) but let's simplify it like that for now.

Ideally you estimate that you should have LTV = 3 x CAC

For Writizzy.com, I only have 6 months of history so my LTV is necessarily limited to 6 months but I still have customers who took annual subscriptions.

Stripe tells me my LTV is 181 euros.

Ok, this might seem weird since I told you I only had 6 months of history and therefore not enough hindsight to have customers who spent this amount.
Stripe actually uses this formula:

$LTV = \frac{ARPU}{Churn}$

• ARPU (Average Revenue Per User) represents the average revenue per user per month.
• Churn is the percentage of customers who cancel each month.

So it's more of a prediction than a reality.

I still think my history will greatly evolve this value. We'll see later but overall it means that already if by spending 100 euros I can sign a customer, then I'm winning.

Reddit Ads

I preferred to start with Reddit Ads to spend my first 100 euros. I tried to use Google Ads but without success and the fact that I didn't have tracking active on my site created lots of errors for me. It's theoretically possible to do without it but Google refused to spend my money :)

Reddit, if you don't know, is a sort of mega forum where people talk about pretty much everything. It's estimated to represent between 1 and 2 billion monthly active users. It's colossal.

Reddit is organized as sub-forums, the subreddits. Each subreddit talks about a topic, for example raclette, people who photograph slices of bread attached to trees… or channels reserved for AI. Ok…

And it's possible to display ads on subreddits, so in theory to well-targeted populations, like, people who like writing blog articles and sending newsletters, which works well for me for Writizzy.

I set up two campaigns to test two approaches:

One campaign emphasizes the price difference with Substack.

The second emphasizes that we must remain masters of our data.

I'll spoil it right away, I changed the ads and actually I focused everything on price after 3 days.

Here's the graph of impressions and the amount spent per day provided by Reddit:

We note:

• that I spent exactly 100 euros, at 20 euros per day for 5 days
• that my number of impressions increased drastically on the 4th day.

Because yes, I made some big mistakes…

Campaign setup

First mistake is that I thought it wasn't working. That is, I started the campaign on June 9. I did see things happening on the 9th and then for 2 days I thought Reddit wasn't doing anything anymore.

In fact each day I would come back with the same date filter so I didn't see more than June 9 and until June 11 I thought Reddit wasn't working…

To be honest, I also thought it wasn't working because I saw barely any additional traffic arriving at my site.

And then I fixed it and traffic arrived. Here's what I changed:

• I switched from a Free form ad type to Image
• I strongly restricted the audience by removing many countries from geographic targeting

The Free form ad looks like a post and then the person has to click on the link placed in the text. Whereas the "Image" ad sends traffic directly to you when people click. Let me tell you the difference is gigantic. You can see it in the traffic, nobody came with the Free-form ad.

Audience restriction has other benefits, two for my case:

• my product uses a payment platform that doesn't allow payments from all countries, so it's pointless to target unsupported countries
• I'm subject to a lot of spam attacks from certain countries, might as well stay quiet about them…

It's not a problem to target fewer people as long as they're the ones ready to convert and pay.

So I reduced the audience to North America and Europe.

Landings and conversion

Starting from June 11, I finally had traffic and we're talking about significant visits, between 5 and 7 times more than my usual traffic. Reddit became my number 1 traffic source at that point:

But it's nice to have visits, but it still needs to be useful.

At first I was sending to the comparison page with Substack. But this page isn't adapted for conversion, it's oriented for SEO. It doesn't match the ad that was on Reddit and which talked about the price difference with Substack.
(Plus here you see the latest version of the comparison page which was worse at that time).

So I quickly made another page just to be a landing page for the Reddit campaign.

The difference between a SEO-oriented page and a landing page is that the second can have conversion-oriented messaging and must limit the number of exit points outside the signup button. Whereas the "SEO" page should on the contrary participate in the internal linking of the site and should be more neutral to hope that the indexing bots propose it in search results.

The bottom line

Over these 5 days, I spent 100 euros, 20 euros per day.

In reality only 2 days were really effective because of my startup errors.

My tracking in openpanel was a bit off and I had to improve it to better see the origin of signups. I can hardly say that the signups from the period are linked to my campaigns. But it will be more the case for the next ones.

I had 1000 unique visitors over 2 days, on a population rather interested in my product. For info, I was doing 2000 unique visitors per month before, so adding 1000 visitors is a 50% increase in traffic. If I had properly configured my campaign from the start I could have aimed for 1500 more visits.

90 people clicked on signup, 44 made it all the way through. That's still a very good number!

I have no paying conversions so far. Sure, a person can decide to convert their account in the future but for now it's not the case. So my CAC is not calculable and I'm not profitable versus my LTV.

A single paying customer would have made the campaign profitable but I can still convert free users in the future.

So will I try again? Yes certainly.

With the same budget I could have doubled or even tripled my monthly traffic if I hadn't made mistakes.

And then it's up to me to better manage my conversion on site.

I'll try to see how I could send the conversion signal to Reddit without sacrificing privacy, if it's possible. Apparently I should look at Reddit's conversion API.

And I'll try other angles for future landings.

It's definitely far from my comfort zone. But actually it's interesting to play with all these parameters and especially to see results. I'm very numbers-oriented so this game is likely to be very addictive on the contrary.