As unbelievable as it looks, the yellow line in the chart below represents impressions and the blue bars are the resultant conversions. We went from getting 45,000 to 1.3m impressions week-on-week in just one account. That’s a 2900% uplift in impressions, which delivered a three-fold improvement in our Twitter Ads ROI performance.
Despite appearances, this isn’t just the result of a quick-win tactical shift. Rather, it’s the manifestation of a change in how we approached the platform, coupled with a strict adherence to tried and trusted best practice. This article walks you through some of the concepts you need to get on top of if you want to start maximising your Twitter Ad investments.
We focused on three main things:
1) Identifying opportunity
2) Understanding engagement, and
3) How to increase engagement rates.
Twitter’s lack of growth (compared to rival platforms) is never far away from the headlines. But don’t let that fool you. The platform’s user base isn’t small, and contains a unique group of audiences.
They don’t behave in a the same ways other social media users do, opting to stick firmly to their platform. Many of Twitter’s 320 million users can’t be reached on Facebook. So if you haven’t yet set up your Twitter Ads campaigns, there’s a fresh set of audiences waiting for you, who most likely haven’t seen your ads before.
Since the initial Twitter Ads launch, MVF has set up more than 50 global @handles, which generate in excess of 60 million impressions a month, meaning at least 20% of Twitter’s global users see one of our ads each month.
There is a great deal of opportunity on Twitter. But you have to know where to look for it. Unlike the ubiquitous Facebook, market leader in most countries, Twitter use varies massively from country to country. Indeed, there are many surprises out there that you may be able to exploit.
The first thing to point out is that 80% of Twitter’s user base resides outside the US. That’s a lot of people. And within that huge majority, the platform enjoys notably high reach in countries that may surprise some. For example would you have expected Indonesia, Saudi Arabia and Turkey to make up the top three in terms of percentages of internet users on Twitter?
Pay attention to factors such as Twitter consumption by country. You need to know everything about internet and social media use in your target territories. There may well be latent opportunity waiting to be exploited.
But it’s not just enough to know that they are using a specific platform. You need to know how they are using it. And when it comes to Twitter, this ‘how’ means mobile. A huge 82% of the platform’s monthly active users’ traffic comes from mobile devices. That means it’s hugely important that your website delivers seamless mobile experiences. There really is no point investing time and money into Twitter, if you’re letting your users down once they decide to click through. When it comes to doing business with you, it’s essential you minimise the barriers your prospects experience.
An engagement on Twitter can come in the guise of likes, retweets, replies, image expansions, home page clicks or link clicks, with the latter usually being the ultimate goal. The thing to remember here is that even though you’re only paying for that link click, it makes sense to get as many people as possible to enagage in some way. Even people that aren’t clicking the link, and have no intention of buying from you, can bring benefit in terms of reach and engagement.
This is important for two reasons. The benefit of increased reach is pretty obvious: more eyes on your ad should mean more conversions, or at least more chance of conversion. But engagement also has an impact on your Ad Score, and this is significant.
Ad score is the result of the algorithm by which Twitter decides who wins ad auctions and how much they have to pay. Twitter wants to encourage brands delivering engaging popular ads, and it therefore gives preferential treatment to organisations capable of delivering along these lines. Here’s how they calculate Ad Scores:
Like most, our initial strategy was to target users who were more likely to engage with our ads/tweets by focusing heavily on targeting. This seemed to make sense, and falls in line with obvious marketing best practices.
But once we started getting comfortable with the system, our focus quickly shifted from targeting small audiences, which were more likely to engage meaningfully with our generic ads, to larger audiences and a focus on adapting ad copy and images to make them more engaging. We saw massive increases in engagement.
If you think this methodology could work for you (and it certainly isn’t appropriate for everyone), make sure you get those people who don’t want the product or service to write back or retweet. Not only will this increase your ad score, but you may even strike gold and see someone with 5000+ followers retweet your promoted tweet. That is essentially an endorsement of your product or service. And what’s more, link clicks from retweets usually get much higher engagement rates, and best of all, it’s free. We’ve seen reach increases of 30 per cent as a result of doing this. It really is a multiple win.
If you manage to adopt this methodology, don’t fall into the complacency trap. Twitter is very much focused on the now and the next, and that means ad fatigue and diminishing returns are particularly prevalent.
The number of impressions garnered and auctions entered naturally drops over the course of a campaign. That means the objective quickly becomes figuring out what do you have to do to maintain a stable level of impressions and spend.
To a certain extent this is beyond your control: engagement rates naturally drop over time. But as soon as the overall engagement rates on a campaign drops below one per cent you start to lose impressions and risk spiralling costs. At this point you have to start considering launching new tweets and new campaigns or even stopping and relaunching your current campaigns.
How to increase your engagement rate
The trick to succeeding with Twitter Ads is to keep your engagement rate as high as possible. If your content isn’t seeing decent levels of engagement you’re not making the most of the platform’s ad offering, and it’s unlikely that you ROI will be as high as it could be. So, to conclude, here are seven Twitter tricks to get those numbers up:
1: Ask questions
This may sound like an obvious one, but it can’t be ignored. Asking questions invites users (impressions) to reply to your tweet and engage.
2: Ask for Retweets
A largely ignored tactic, but when you ask for a retweet you have a good chance of getting one. Salesforce has reported it results in 12 times more retweets. Don’t ask, don’t get.
3: Use Hashtags
Although this can seem like a distraction from your objective (especially if it’s a link click), hashtags have been known to double engagement rate, and they create higher organic visibility on Twitter. Don’t go overboard, though. Our data suggests using more than two hashtags in one tweet can decrease engagement rates by up to 15 per cent.
4: Use Twitter polls
Twitter polls secure really strong engagement rates. In the tweet below, we simply asked the target audience who they prefer out of three singers and we had almost 20,000 votes.
5: Upload multiple images
By attaching multiple images to your tweets, you encourage users to enlarge the images and swipe through, which - of course - counts as an engagement.
6: Use UGC
User generated content fuels the internet, and advertisers around the world are seeing huge uplift in CTR and engagement rates by incorporating it into their content. We’ve found that the more natural, and almost unprofessional, the image looks, the better it does.
7: Schedule for maximum engagement
There are many articles all over the internet that suggest Saturdays and Sundays are the top performing days in terms of Twitter activity, though this of course feels like it leans slightly more in the direction of a consumer campaign. While our data does provide some evidence for the theory, it’s important to remember that industry averages and benchmarks mean less than your own historical data, so that’s where you have to focus.