How we unnecessarily burned 20k with Facebook Ads

We are in the midst of 2014. The first months after the Beta-Phase had passed and our first major update brought many new features and a completely new design. Additionally and shortly before that, we had presented Stackfield at different trade fairs, such as CeBIT or Tools, which generated a very good response to our product. Encouraged by these events and successes, we took the decision that it was finally time for a large-scale advertising campaign. While offering a pure online tool, the choice fell naturally on online advertising, especially that we had previously conducted a small advertising campaign in the form of sending letters to agencies that proved unsuccessful – the road from the offline world to an online tool was simply too long and our mailing list was of such poor quality that a large part of the letters was returned.

At that time, advertising on Facebook was still relatively new, but we anticipated a great potential here: whether a writer at a marketing agency, a managing director of an IT company or a lawyer at a major law firm – almost everyone had a Facebook account and could be reached through targeted campaigns. Compared to the alternatives, such as banner advertising, Direct Mailing, Google AdWords, etc. we were thus expecting a great success.

The Implementation

Getting started with Facebook-Marketing through the Facebook Ads editor is extremely easy, which is why we started to create ads right away without major detours. With this tool the upload of multiple images in order to test some Sets, the definition of different target groups and also the setting of geographical boundaries are straightforward.

And so the first campaign was created quickly, but we soon realized that there were some problems: although we could see in Google Analytics how many new visitors came to our site from the campaign and how many registered, there was no possibility for long-term tracking. Additionally, it was quite clear that there was a vast amount of fake accounts on Facebook. The number of likes on our Facebook page grew disproportionately fast compared to the actual visits and registrations on Stackfield.

The latter problem is considered a waste and unfortunately it must be accepted when using Facebook.

We solved the first problem through a separate tool, which labels the user with the originating source when registering. This allowed us to continuously see which campaigns and sources were actually successful and which weren’t. The derived results were anything but pleasant. Let’s consider the period from 11.08.2014 to 11.11.2014 (Facebook’s Fanpage tracking cannot go back more in time, while Google Analytics cannot show more than 93 days …) as a reference:

We spent exactly 18.428,99 € on 10 different campaigns, each with its own variations. Through them, we have reached 2.262.403 Facebook users. „Reach“, by Facebook’s definition, includes the users that have potentially noticed the ad (e.g. the ad was displayed) – no interaction with the ad is required. On the other hand, the ultimate goal of the users actually visiting our website was met exactly 31,600 times. This results in a CPC (Cost-Per-Click) of 0,58 €.

Surprisingly, Google Analytics reported only 28.818 visits by users who were redirected from Facebook during this timeframe – i.e. 8,81 percent less. From these users, 3.786 registered on Stackfield.

Up to this point, the results from Facebook-Marketing proved to be surprisingly good. Even more: these figures were considerably better than the ones obtained by any other marketing channel we were using. We had spent € 4.87 per new user. Back then, like today, this would have been a more than acceptable cost for a new user.

As already mentioned, there were certain circumstances that spoiled the joy of these results. Our proprietary tool that tracked the source of our users, yielded a clear conclusion: the churn rate among Facebook users was enormous and at the opposite pole when compared to the other sources – and this at a time when the tool was still free to use. The Facebook users were relatively cheap and quick to acquire, remained for a short time and were soon gone – and nearly all of them.

We considered various causes for this problem and tried to solve it through better targeting, more specifically by further optimizing our settings. Targeted ads for counselors or agencies, which increased fivefold the CPC while bringing the right people, did not alter in any way the previous results. Even these users churned after a short time. Why is that? We do not really know! Eventually, Facebook may just be the wrong audience for a B2B tool. Most people visit Facebook in their free time and are not actively looking for a tool like ours. And that seems to generate the different conversion rates: users who come from sources as Google or through referrals remain on Stackfield.