Should You Use Auto Direct Messages on Twitter? The People Science, Research, and Perceptions Behind Them.

On Saturday the 19th of March, I decided to stop messing around and grow my Twitter audience.  You see, I was there, I had been on Twitter for a while but I had never made any attempt before to grow my audience or if I’m honest, to even engage properly.  That changed on the 19th of March and this is the first in a series of blogs that looks at the science and perceptions of growing a Twitter audience. 

Today is all about the auto DM.

The Auto DM

On Saturday the 19th of March I had 2,505 followers and today on the 27th of April I have 4,733 followers.  I did several things to grow this audience but we will cover that in a later blog, today is about the large volume of auto DM’s I received during this growth.  In total I received 628 auto DM’s, equating to 14% of the people I followed. 

14% of the people I followed use auto DM’s in the first step of engagement.  Ok, the last part of that sentence is a little strong, not everyone’s auto DM is really about engagement.  I also thought that was quite a large number of people opting to use them. 

So, naturally, I decided to test out my own auto DM’s to find out if they work and what people think about them (more on that later).

The use of auto DM’s has been about for a long time and for the same amount of time there have been warnings [and massive rants] about their use.  Yet, in my perception, the number of people using them has increased rather than decreased.  So, before we get to the research let’s look at the ‘science’ behind them.

People Science of Auto DM’s

I’m struggling to write here… Really, I am.

OK, the use of auto DM started for two reasons:

1. Because the functionality was made available
2. Because of the desire to automate as much of social media engagement as possible

It doesn’t exactly scream social.  The practice is linked to spam and from the many, many blogs on the subject, they appear to really upset people.  But maybe they are being used in the wrong way (my optimism shining through), let’s take a look at the assumed science:

It has been openly declared for many years that social media should be through of as relationship building.  As Ted Bauer says all marketing activity can be attributed to building relationships.  The Auto DM is the fist nod, the first hello, in building a relationship but the automation comes at a price.

Many view the automated ‘hello’ as inauthentic; there is a desire to take the time to say hello as a person, not a bot.  There have been many studies on social media connectedness and the social brain (a book I love, Social: Why Our Brains Are Wired to Connect), an automated response stops this connection as it is not a person behind the message.

Social media engagement isn’t about you, it’s about your audience.  There is just too much noise and too many choices out there to be heard if you don’t know how to add value to your audience.  Twitter is a great place to test out assumptions but you do need to test and refine, take this small piece of research on Auto DM’s as an example. 

I hope we’re in agreement that auto DM’s appear on the surface to be the first step in relationship building.  But they lack authenticity and most probably an understanding of your audiences’ needs.

Let’s test these assumptions.

The Research

I sent 1,292 auto DM’s during my growth.  I used two auto DM’s, one that was more in line with the DM’s that I was receiving and one that came from my desire to understand more about peoples’ perception of auto DM’s.  I also had a ‘control’ group where no auto DM was sent. 

From the DM’s that I send 22% were from the tweet reminiscent of the ones I received, 36% were from the one asking opinions about auto DM’s.  42% of the people who followed me had no auto DM. (I know this wasn’t even but I had little control over the automation of what was sent, scary).

DM #1:  I received ZERO replies to the first DM, there were no clicks on the link either.  I’ve checked and while no one has responded, they didn’t immediately unfollowed me either. I ended up stopping this auto DM early because there was no engagement at all and I was concerned that it was impacting my perception.

DM #2: I received 9 replies to the second DM, a total of 1.1%.  These were more than one-off conversations, they continued over a period of time and these people were happy to be included in my research (hi guys, you made it!).  2.3% of people who got this auto DM clicked on the link.  Again there was no immediate unfollow. 

I responded to all these tweets, all 9 of them, but in other research I’ve read (non-scientific research) they found that no-one replied to any response from an auto DM.

Eeek.

The ‘science’ behind auto DM’s is to build a relationship.  If you’re not responding, then you’re not even attempting to build a relationship.  There is no science, no persuasion behind what you’re trying to achieve, we’re back to spam. 

This reminds me of a conversation that I had a couple of weeks ago about reduced Twitter engagement and lead generation.  The conversation was about the desire to increase lead generation on Twitter.  The group I was speaking to didn’t want any brand awareness, or really to increase engagement they just wanted to increase leads.

Let’s think about this for a second… Do you think that is going to work? In my opinion, it is this desire that breeds the appeal to use auto DM’s but I’ve just started to prove they don’t work.

In fact, when I was telling my friend Sean about this research, his response was:

“Auto DM’s are just a bit needy… stop trying to sell me your shit, it’s too early in the relationship for that”. 

I tend to agree.

No DM:  I received replied and engagement from the people that received no auto DM.  In fact, I received a higher rate of engagement from the people that received no immediate engagement from me.  This is partly down to the content that is posted on my Twitter account.  I also have a pinned tweet that converts well, this tweet is a bit provocative and get’s people talking to me. 

Here’s the thing about behaviour, it’s not logical.  In the auto DM, it’s like we’re listening, a little too hard to behavioural economics and thinking that people are too lazy to check out the profiles of the people who follow them, and instead just spam them with content. 

In fact, people do check out your profile.  In my research, I increased my profile views by 1084%, in 41 days.  Clearly I’m not for everyone as I’m converting about 20% of this number into followers (assuming that the visits are not from people who already follow me and this statistic is driven by my growth strategy – Twitter that would be a great metric to measure!) but clearly people are looking at my profile. 

From this process, my advice would be to let people explore you and decide if your content speaks to them.  Take the time to build the relationship, don’t think about the immediate lead generation like the group above as you will probably go down the wrong path with your engagement tactics.  I write a lot about the functionality of social tools and the thing I continually find that is that just because the functionality is there doesn’t make it useful.

I know these auto DM’s come in all shapes and sizes, I received 628 of them in 41 days.  Of course, I did a bit of research on these too and I’ve created an auto DM typology.  If you use auto DM’s see where yours fits and the perception you are creating.

Typology of Auto DM’s

I’ve found that there are 7 main types of auto DM, there are also others that have a few of these elements mixed together – talk about multiple CTA’s, come on people!

#1 Visit My Blog/Website

At 26%, asking new followers to visit a blog or website has the highest frequency of use.  The messages differed in tone and perceived forcefulness of visiting the external link but the tactic behind the message is to move the audience onto owned property. 

Given the low click through rate of my own Auto DM’s I don’t see this type of post converting well.  It also begs the question whether the people sending these are checking the conversion rate or whether people re-engage with them. 

#2 Connect on Another Channel

Attempting to drive a new audience to connect on other channels including LinkedIn, Facebook and Instagram was they second highest type of Auto DM at 21%.

Again, as my friend Sean would say this feels a bit needy "Please be my friend in Facebook".  Interestingly, in my research, I found that the people really interested in you will track you down on other platforms anyway.  I drove traffic to all my other social profiles, growing LinkedIn by 10% directly from Twitter.  There is no need to be perceived as needy, instead, concentrate on providing good content and value to make your audience more interested in connecting in other channels.

#3 Free Giveaway

So, we’ve probably all heard that developing a lead magnet or giveaway to collect email addresses converts. It may stand to reason that it would convert on Twitter too, 19% of the Auto DM’s I received were offering some kind of ‘value added giveaway’.  I didn’t measure this but with the strong dislike to Auto DM’s I can’t imagine that they convert well.

I found that the tweet I had pinned to the top of my profile converted and drove discussion about it, there was no need to Auto DM people.  It is still in a prominent position on my profile and when people looked at my profile it was the first thing they noticed.  It may be an idea to see if a pinned tweet converted better than an Auto DM, some testing to consider.

#4 Thanks for Following

The fourth most popular Auto DM was a simple ‘thanks for following’ at 18%. 

Again this was not a metric that I tracked in my research but the feedback I was given is that even the thank you DM’s are annoying.  I know there is a desire to try and ‘connect’ with new followers but the feeling is that this connection should not be automated.  Automation does exactly provoke feelings of authenticity in others! 

#5 Ask a Question

Asking questions to new followers accounted for 10% of all Auto DM’s I received.  I certainly asked a question in my research but the tone and the questions asked from others I received seemed a little forceful. 

I did find that asking a question converted at a low 1% but is it really worth annoying new followers for a 1% conversion?  I found from my control group that if people like what I was saying they would find the blog themselves and sign-up.  I think some of the people I spoke to by asking a question were surprised when I answered.  If you do insist on doing this, then please do respond!

#6 Direct Sell

I have to say I was a little shocked about this but there were some tweets that had a feeling of being sold to directly.  These types of tweet accounted for 3% of all Auto DM’s I received

Needless to say, these people were unfollowed very soon after.  My feelings on direct selling are that if you are trying to sell to me in the first engagement then it’s only downhill from there and I really don’t want to be sold to like that.

#7 Can I Help You

Finally, the 7th type of Auto DM I found was the ‘can I help you?’.  This also equated to 3% of all Auto DM’s I received and while appearing a little more genuine on the surface, when reading these tweets, I feel they are a little disingenuous and self-serving.

Again, this was not an Auto DM that I shared so cannot comment on the return I received but from the ones I received I do not see an immediate motivation to respond.

Moral of the Auto DM Story

Don’t do it! 

I’ve now turned my Auto DM’s off and I’m still receiving inbound conversations and increased levels of engagement.  In my experience, they don’t convert, and they do piss people off!

I get that you want to reach out, I get that you want to engage and tell people about all the amazing things you do.  The trouble is, it’s too early and intrusive to your audience.  Your new audience is not that lazy, they will look at your profile if you concentrate on producing value for your audience. 

However, this will only come from a deep understanding of who your audience is.  During my growth process, I’ve learned a lot about who I appeal to the most and what value they are looking for.  Are you doing this research in the background of your engagement?  It is this willingness to provide value to your audience that will generate leads, not the auto DM.

What do you think about auto DM’s?  Have you done any research on them?  I’d love to hear your thoughts.

Connect Deeper

Dr Jillian Ney is the UK’s fist Dr of Social Media and a Digital Behavioural Scientist.  She uses people science for digital data to help businesses understand what’s driving their customer behaviour.  If this content resonated with you, subscribe to her personal blog.