Update on our CSA deliverability comparison

We revised the presentation of the blog post shortly after publication — even before the first warnings were received — and then, in particular, removed the original ranking rating, so that the comparative character based on the tested criteria and results came into focus. In addition, we had to correct three manual mistakes that we had made during a total of 480 individual tests and which had been complained about in two of the five warnings, among other things.

At the same time, ESPs came forward who saw the gaps we identified in their implementation of deliverability criteria as constructive criticism and set about readjusting their account settings and technical parameters. For example, Jens Klüsener, Head of Marketing at Cleverreach, informed us on May 14 that after publishing our blog post, his company had incorporated the previously missing DKIM functionality and activated it for all customers. In addition, the list-ID header entry was implemented system-wide. DKIM was able to activate Cleverreach for all customers in one go because their email platform does not work with customer-specific link domains, meaning that customers did not have to make individual name server entries for DKIM.

However, an ESP was not satisfied with the adjustments and corrections we made, but wanted to continue to see the blog post completely taken offline. He therefore applied for a preliminary injunction with the Munich Regional Court I (Ref. 33 O 3751/14) in February 2014 and was therefore initially successful as part of the urgent decision, which was issued without oral hearing. That's why we had to temporarily remove our blog post from the web. However, we filed an appeal against the decision and were thus able to assert ourselves after oral proceedings, so that we were able to put the blog post online again on April 30. On the most important point of contention, the question of whether a total of three emails per ESP are sufficient for a settlement, the district court ruled in its reasoning:

(...) is the random analysis of three emails from different customers with regard to the overall range of the platforms tested.

The judgment was accepted by the plaintiff ESP. He did not appeal the judgment and did not bring any main factual action because of it.

Even the Certified Senders Alliance, which we in another Blog post about our CSA deliverability comparison had carried out an overly lax audit of ESPs with regard to compliance with their criteria, responded to our settlement. In a conversation at Email-Expo in Frankfurt, Ivo Ivanov, Director of CSA, and Sascha Wilms from the ISP & Product Development division complained that we had not appreciated the actual focus of their work: Around 80 percent of their working time went into managing complaints received from users or ISP sides. According to the two CSA representatives, this process makes a significant contribution to motivating senders to comply with ECO guidelines and deliverability settings that comply with the guidelines.

However, we never intended to conceal the CSA's complaint management. However, this point was not the subject of our comparison. This is because the complaints are usually directed against (presumed) unjustified email submissions, for example against the background of a lack of consent to receive advertising communication. At best, this has something to do with the technical criteria tested by us, which the CSA makes a requirement for participation in its whitelisting project.

With our CSA deliverability comparison, we are pleased to have also initiated a further debate about the utility value of association certificates. This is how iBusiness author Dominik Grollmann diagnoses under the title “Alles nur Lägen? What service provider certificates are really good for” on March 19:

Although the (...) Certified Senders Alliance (...) also issues' certificates' — in fact, new members essentially sign a declaration of commitment. The CSA only reserves the right to carry out an unspecified “sample check on the fly” and also operates a complaints committee. There can therefore be no talk of a real certification process. The term “certificate” is simply misused.

The basic conflict is thus reflected, as is so often the case, in the revenue model: Industry associations and the seals of approval enshrined in them usually live on the often generous contributions they collect from their members or participants. If a seal of approval is part of the association's “membership package” or can “buy” more than is earned as a paid add-on, there is therefore a risk that it will say more about the financial capabilities of its institution than about the qualification it is intended to prove.