Measuring the performance of your digital channels

Digital marketing is cheap. But is it effective? The sad truth of the matter is that many VCSE organisations may not know. While stakeholders demand a consistent flow of numbers in an attempt to evaluate projects, very often we don’t measure the impact or seek to improve those things which don’t cost us any money.

While we may undertake some rudimentary reporting, not having a full grasp of the figures ensures that the net result is that many charity social media accounts, email newsletters and websites aren’t performing as well as they could be.

Treating your digital channels as an extension of your traditional marketing activities is common, but not ideal. Your audience today expects your online presence to service a need. You can only understand and respond to these needs if you begin to measure how your audience behaves and responds to what it consumes online.

1. What to track

Online analytics that tend to get reported on are reflective of a time when people didn’t really understand digital. Typical of these will be ‘number of followers’ or ‘number of pageviews’. This is great but it tells you very little about how you can improve channel performance.

For your social media you should be concerned with two main measures – impressions and engagement. The former will tell you how widely your posts are seen while the second will indicate how well they chime with your audience and how they’re contributing to your overall objectives. With email you should be focussing on open rates and click rates.

2. How to track it

The first step should be to measure the performance of your channels are they are now and establish how you’re doing, on average. Your best practice lies anywhere that’s better than average so you really need to work out what it is about these posts or emails that made them better than normal.

This is easily done. Facebook, for example, allows you to export a quarters’ worth of analytics via the ‘Insights’ button on your page.  Twitter Analytics can be found via the dropdown menu that occurs when you click your logo. You can typically export data relating to 90 days.

You should now categorise posts by their content and other variables such as whether anyone was tagged, if an image or video was used or if you utilised a link or hashtag. This may be time consuming – but it’s worth it.  

What you should be able to do now is sort your posts by the number of impressions they make and their engagement rate to see what performs best. Work out the percentage of tweets in each category and compare it to the percentage of tweets in your top 50 (or top 100 if you have loads). You can do this for category, day and time.

What you should be left with is a fair indication of when your posts perform best and what difference other variables make.

3. Testing

You should by now be using a social media tool to schedule your tweets. If not then you should, if only because it’s far easier to put a block of time aside every week to write and schedule your online content. It will also make it easier to act upon what you find. Schedule your posts for the times, days and formats that work best

You should notice an automatic uplift in your visibility. Your engagement rate may take a slight hit but don’t worry - you will have more engagements overall. We can work on engagement rate later.

4. Email

With email the same logic applies. Find out what your average open and click rates are and aim to raise them. However, if you do a newsletter that goes out at a regular time and have nothing to compare it with then it’s worth A/B testing it. Try it first with the times of the day, and with days of the week. It’s important to ensure however that the A of the test should be consistent.

5. What did we find?

About eight months ago NICVA started to have a good look at our digital channels. We carried out the process described. Getting under the bonnet told us that:

• Our Facebook, Twitter, email bulletin and website audiences are subtly different.
• Our Facebook and Twitter audiences responded differently to different types of post
• With regards sending emails and posting on social media, there were optimal times, sub-optimal times and times when it really wasn’t worth doing anything at all
• Visability and open rates were primarily affected by timing. Engagement and click through rates were primarily affected by content style and design. Tuesdays and Fridays are bad days to share content. Items posted after 4pm were seen more and engaged with more.

This enabled us to better plan and schedule our content around the needs of our audience. As a result we achieved a 66% increase in our Twitter visability and an increase in engagements of 59%. Our email open rates went up by 30%.

6. Being engaging

Working out how images and hashtags affect engagement is just the start. You can go into detail with A/B testing. Try posting the same content on the same day and time (obviously in different weeks) with some subtle changes. These may include the colour of your images,  the use of emojis and the placement of links your text. Monitor the engagement rate only and see what works.

In email this is best done through the A/B testing capability of your email provider. Many providers give you the ability to test two different types of email content. This enables you to test how you write the content and how you present links to your website for example.

By doing this NICVA found that:

• Content preferences vary – for example, the audience for our ScopeNI newsletter preferred to click on big red buttons that said ‘read more’ and our eNews audience liked to click on hyperlinks in the text.
• Unless we were running or participating in an event, the use of hashtags made no difference
• Posts posed as questions were more engaging
• People engage more with posts with emojis in them
• If you put a link towards the middle of a post, rather than at the end, more people will click on it
• The optimal Facebook post had between 90 and 140 characters

Note: If you attempt to emulate the above you may end up with completely different results because everybody’s audience is different.

7. Conclusion

While the immediate benefits of having your brand seen by more people and having more people click through to your website should be clear, the longer term benefits of being attuned to your user’s needs will be priceless.

You should combine this information with that found in the ‘demographics’ section of your social media analytics, and Google Analytics to better understand the characteristics and behaviours of your audience.

Then you can look at segmenting it. More on that later…

8. Some useful tools and resources 

• Mailchimp
• Twitter Analytics (requires you to be logged in)
• Facebook Analytics