Identifying engaged subscribers: unique opens, clicks, lateral thinking
Latest posts | Feed | | By Mark Brownlow
Last week we saw why it's important to identify your engaged or "best" subscribers. And we discussed how repeat opens is one way of doing this when you don't have funky customer data at your disposal.But what other techniques can we use?
In this post, we'll cover unique opens, clickers and how we might apply lateral thinking to come up with other neat measures that don't require much work or data.
Unique opens
Your average subscriber does not open every email from you. So if you're looking for engaged subscribers, how about searching for those who open more than the average?
As in the previous post, I took four months of emails (nine issues of my e-newsletter) and used simple campaign reports to evaluate open patterns. About 68% of the list recorded at least one open over that period.
Of those 68%, this graph shows how many opened one, two, three etc. of those emails:
We see that 35% of our openers opened at least half of those emails. And 8% opened every single one. So we might consider those 8% to be our best subscribers.
There are, of course, two problems with this approach.
First, a single open does not imply anyone actually "engaged" with the email, which is why we looked at repeat opens last time. A recipient might click on every email, trigger an open in a preview pane and delete it without reading.
That would put them in the 8% but you'd hardly call them engaged.
Equally, there are those reading the emails religiously who never trigger an open because they use an email client where images are blocked.
Fortunately, even basic email campaign reports also tell you who clicks on an email. So instead of looking at unique opens, perhaps we should look at unique clicks. After all, if someone clicked on a link then they must be interacting with the email.
Clickers
A look at the campaign stats for that B2B newsletter of mine found about 40% of the list had clicked on at least one of those nine emails.
So here's the same graph, again. This time it records how many of those 40% clicked on one, two, three etc. of those emails:
If you compare the two graphs, you see why it's so important to look at numbers that represent genuine interaction. About 25% of the entire list opened at least half of those nine emails. But only 3.5% of the entire list actively clicked on over half of those emails.
Still, we can treat those 3.5% as genuinely engaged subscribers.
Apply lateral thinking
If these open and click patterns aren't enough, how about digging deeper?
Example 1: If someone really likes your emails, they would maybe keep them and refer back to them later.
How about looking at open times or click times for past emails and isolating those subscribers who opened or clicked weeks and months after the email went out?
Example 2: We know that the longer someone is on a list, the less likely they are to open and click on emails.
So someone who opens every second email and signed up in 1997 is likely far more engaged that someone who opens every second email and signed up last month.
So how about splitting subscribers into segments based on sign-up date, calculating average engagement metrics for each segment and then picking out those who engage well above-average in each segment?
Example 3: Are all clicks equal? Are there some links that suggest only those with a close relationship to your brand would ever bother clicking? Are there links where you can immediately tag anyone clicking them as a "top" subscriber?
What about "behind-the-scenes" links like "learn more about the team" or "see how our factory works"?
What about people who click on customer survey links and complete the survey without any kind of incentive?
Can you find links that identify engaged subscribers in that subtle way.
My newsletter has the title "No man is an iland" and an obscure link nearby that reads "Why no S?":
I might decide that anybody wanting to know the origins of the newsletter's title is likely more than averagely interested in what I have to say. Incidentally, 2.66% of the list have followed that link.
What other measures can you find to "out" those best subscribers?
The ultimate subscriber
If you're not happy with one or another measure, you can use two or three, compare the addresses identified and see if there's any overlap. Anyone who comes up as "engaged" using each of several different measures really is a "best subscriber".
I did the exercise and found one subscriber who fits all these criteria:
- Actively subscribed for over two years
- Clicked on more than half of the last nine emails
- Registered an open on each of the last nine emails
- Clicked at least once on the "Why no S?" link
- Recorded more than five times the average number of total opens on at least one email
- Recorded more than twice the average number of total opens on at least two emails
(And, no, it wasn't me.)
Final thought
Many of these measures of engagement are biased by those who signed up sometime over the period of comparison.
Sometimes these new subscribers are unfairly penalized: it's hard to click on more than one email when you've only ever received one email.
Equally, new subscribers tend to open and click more often than old timers. So they can appear more engaged than they truly are.
This high level of initial engagement is why it pays to send new subscribers a specific stream of welcome messages that take account of this factor.
It may also make sense to keep new subscribers out of your "top subscribers" pot until they prove their continued interest after the initial novelty of your emails wears off.
Permalink | July 17, 2009 | 6 comment(s)
Get posts like this: as an RSS feed | biweekly email | via Twitter
6 Comments:
Thanks Mark for another great detailed post.
I did occur to me however that i was reading the post, as i do all your posts from my RSS feed. As i want to get your posts as soon as they are published ( yes i am that engaged) by the time i receive your newsletter I have alreadly read all your posts, and generally just scan to make sure i havent missed anything-- so i guess i wont show up much on your stats. which i guess is an issue - how do we measure our engagement and impact by communicating on social media tools such as twitter, RSS, facebook and the like? As i now enagage with you in some many ways, does this make the email metric, less reliable - or do we/can we intergrate our metrics with social media engagement?
Thanks
Claire
By 
Rollo, on
23 July, 2009
Hi Claire! That's a brilliant question which goes to the heart of many issues surrounding integration of online marketing channels, especially ones like RSS that are harder to measure.
I'm still getting my head round these issues, but two points that leap to mind (apologise for brevity, but I'm on holiday at the moment):
1. Should we abandon the idea of thinking of people as email subscribers, twitter followers etc.? Do we care where they interact as long as they interact?
A holistic approach looks at the overall picture and invests time and resources where it makes most sense without any false notions of one channel being better than the next etc.
(I'd encourage people to think of themselves as marketers who use email, rather than email marketers. Once you have a label, you start to lose objectivity and can get unnecessarily defensive.)
But it's hard to implement, because you then have to think very hard about the indirect benefits of each channel. It's relatively easy to measure clicks from an email or through a short url at twitter. But if you go holistic, you also need to know this:
If I didn't send Claire an email, would she still stay as engaged through twitter/RSS? The email metrics say you don't click, so losing email would not be a problem. But as you explain, the email is still part of the whole interaction. You do look at it. It's another touch point, another reminder of this Brownlow guy's articles. So it has a value, but it's hard to measure.
And I think the technology for a truly holistic approach (where data is recorded for customers across all channels) is accessible only to a very few right now (if anybody). Ideally, I'd like to identify the people who read the blog, get the tweets and read the email. But how can you link a twitter follower to a blog subscriber to an email reader? That's very hard to do in an automated way.
So I strongly believe you have to follow your instincts to some extent.
The second point is how you see those different channels. Are they alternatives for the same message or do they offer different value?
(I tackled some of those issues here.)
For example, my Twitter channel is different to the blog. Although it promotes the blog articles, it also points people at other places of interest. So it has its own inherent value.
The email newsletter more or less just summarises the blog posts. So the independent value of the email is relatively low, IF you already get the blog feed.
That's where the decisions start. Many people aren't getting the feed, so they are very engaged with the email. So is it worth investing more resources in the email to make it interesting for those who also get the feed? How many read both?
Broadly speaking, I think it's good to try and give each channel a common thread for reinforcing messages, but also something unique so that there is inherent value in interacting through more than one channel.
A major benefit of that approach is that you can then look at the engagement numbers for each channel realistically...
If the email is just repackaged web feed, is a non-clicker not clicking because the articles are boring or because they already got the feed?
If the email has some unique content in there, then the "also reading the feed" excuse does not apply. Poor content (for that reader) explains the lack of clicks.
There are more questions than answers here for now. Sorry!
By 
, on
23 July, 2009
Thanks for the swift and full reply Mark, especially as your on holiday!
Indeed as a marketer who uses direct mail, email marketing, social media and good old display advertising it can be difficult to 'truely' assign ROI on each discrete campaign. I agree we should take a holistic approach, but often SMT want to know ROI on each channel - often there are different people who communicate in each channel which just adds to problems in large orgs- and especially in terms of social media use. Trying to convince senior managers of the value of social media when some of the traditional marketing metrics cannt be applied here can be an issue.
I agree there will always be a need/use for email..i guess for me its providing the communication in the format, at the time the customer/subscriber wants it...so by blogging, sending a monthly newsletter, using RSS, twitter and the like the customer can engage with you in the communication channel THEY feel comfortable in at the time they want...which i guess is main goal when sending our emails - hence why we are always asking when is the best time/day to send our emails?
Like you say the more touchpoints we provide, and the more opportunties we give to engage, this ultimately has to to help increase traffic to our websites and email sign ups.
By Rollo, on
23 July, 2009
Great comments Claire. I'm on your side wrt letting people choose the communication channel they feel most comfortable with (as long as you can support the channel adequately). I didn't always think like that but the more I see and experience, the more I find the world is going that way...ever more power to the end user.
By , on
24 July, 2009
Hi, I too have a lot of questions about how RSS feeds and Social affect "engagement". One one hand I think that for businesses (not just blogging where blogger would be great) with their own websites benefit greatly from using a good cms (like here: http://www.oshyn.com/_blog/Sitecore/post/The_Next_Evolution_of_Web_Content_Management_Technologies/ ) But I also wonder if we can analyze the effectiveness of our emails etc by looking at both our open rates, increases in traffic and conversion rates. I think all of these depend on the business. But I think that no matter how great your emails your website has got to back-up the offering and vice versa. I think many marketers have forgotten the concept of "old school" advertising that it takes on average 20 impressions of a brand for a consumer to react. I might get an email and read it...but I'm pre-occupied or I can't actually buy right now. But if the content is good/interesting etc...I will likely engage.
By kimberly e, on
04 August, 2009
Thanks for the comment Kimberly. That speaks to an important issue, especially in a tough economy. Offers, offers, offers can get tiresome, so what else can you put in your retail emails that keeps subscribers engaged until they are ready to spend.
By , on
05 August, 2009
Comments closed during migration to a new blog platform in early May
