To put it simply, the SendForensics Deliverability Score is a measure of an email's 'quality'. This ‘quality' is in terms of how it will be perceived by global filtering-systems/ISPs and, to an extent, human recipients themselves.
The score itself represents the percentage of inboxes in an average mailing list that an email will reach. It is calculated by modelling the email's acceptance against the behavioural spread of the worlds spam-filtering-systems/ISPs at that given point in time (the environment fluctuates constantly).
There are three assumptions that have to be made to correctly relate the score to an email/campaign:
Mailing list is clean and fully-verified i.e. no bounces
Mailing list is a globally ‘average list’ in terms of its list-composition. This is a list that reflects the broad distribution of filtering-systems/ISPs worldwide (see below).
Mailing list is a globally ‘average list’ in terms of its prior engagement history i.e. from addresses that have never been sent to before from that domain/IP, to those having a long history of engagement with that sender.
The score is accurate if these 'lab' conditions are met, but of course in practice each of these conditions will vary by differing amounts. Crucially, however, following the system's advice for fixing issues and improving the score will improve results under any conditions.
In practice, each mailing list has a different composition based on its audience. For example, a B2C list may have a significant percentage of its addresses with freemail providers (see fig.1), whereas a B2B list will likely have more corporate-owned addresses behind enterprise spam-filtering-systems (Symantec, Barracuda etc). They will therefore behave slightly differently in deliverability terms.
This means the various issues reported by the system will have slightly less or more impact depending on the list but, as mentioned above, there will still be an impact. For example, an issue that may cause a 2% response-drop for a 100% Gmail list, might only cause a 1% response-drop for a 50% Yahoo and 50% Yandex list. But there will be a drop nonetheless, and the advice given in the system can recover that loss before the campaign is sent.
It would be practically impossible to ask every single campaign recipient whether an email reached their inbox, which is why the deliverability metric must be modelled in this unique way. However, what we can do for proof of efficacy is to look further down the chain of engagement to measure the resultant effects. Ultimately, that means clicks.
Deliverability Score vs Unique Clicks:
Real customer campaign results over a one-month period, pulled directly from the Email Deliverability Suite (with ESP plugin enabled)
In order to minimise fluctuations in other metrics and isolate the effect of the Deliverability Score on engagement, the above newsletter was chosen for the following reasons:
Consistent offering (weekly listing of items for sale)
Consistent mailing-list size (55,400 contacts +- 250)
As can be seen, the deliverability score has a direct effect on the number of unique clicks.
Whereas CTR/Unique-Clicks may be the most reliable performance indicator in email marketing, there is another important metric that benefits from higher quality emails: Spam Complaints. Too many complaints can taint your mailing reputation and affect the success of future campaigns.
Adding spam-complaint/abuse data to the same campaign dataset produces a startling pattern when the y2-axis is plotted logarithmically (for data-range comparison).
As can be seen, there is an almost perfect inverse-correlation between the Deliverability Score and number of spam-complaints registered.
The SendForensics Deliverability score is modelled in realtime on the behaviour of a comprehensive distribution of global receiving systems
Specific optimisations and advice impact individual systems by variable amounts, but affect them nevertheless
Ultimately, a higher SendForensics Deliverability Score = more clicks, less complaints
There are many variables that affect campaign success, with some rather more difficult to control than others. Now that it's possible to keep deliverability regularly in the sweet-spot, and pre-emptively before a campaign is sent, that's one less problem to worry about.