Working in paid media, we’re collecting large amounts of data every day. We can see how different audiences interact with our ads, how they behave on our website, how long they stay, what and when they buy and where they go. But how do we generate powerful insights that drive action?
The key to drawing the right conclusions from your data sets is segmentation. Common criteria used for segmentation include: demographics, locations, interests and status in their user journey/decision making process.
The first step to analysing a users journey is the segmentation into users who know you and have recently interacted with your brand (e.g. remarketing audiences, users who have been searching for your brand or direct traffic.) and users who have been targeted through prospecting.
You can also segment your audience by relevant technical factors. One example of this is splitting your target audience into mobile vs. desktop users. Mobile and desktop users will not only experience your website differently, but they might also display different browsing behaviour and involvement.
Having said that, segmentation options are endless and it is important to focus on the most powerful ones. You want your segmentation to be detailed enough to draw insightful conclusions, but also large enough to be statistically significant.
When you are segmenting your data, you will also need to consider the timeframe you use in your data set. For example, you may only look at leads captured in the past month, or you may look at leads captured in the past 12 months. Choose the timeframes you’re reviewing wisely. Changes in your market or in your marketing activities during your chosen timeframe could affect your results. If possible, consider not only online but also offline, above-the-line activities.
Having segmented the data will enable you to identify trends or patterns within or across segments. When looking at your chosen time frame, remember to consider possible seasonal trends, market trends and campaign-influenced trends.
Finding the same trend across different channels or traffic sources indicates either a market trend, your market competitiveness or technical issues with your website(such as loading time or issues with the checkout process)
If you want to know more about how your offer compares to competitors’ offers, your brand campaign or direct traffic is usually a good indicator. A drop in conversion-rates on your brand adverts or direct traffic could mean that your competitors have a stronger offer. On the other hand, if you see brand conversion-rates remain stable, but all of your non-brand campaigns are dropping, you might need to review your campaign strategy.
Patterns are changes in performance or events which happen repeatedly and in a predictable manner. These can be regular and scheduled events, such as seasonal trends, sales periods, trade fairs, public holidays and changes in performance on certain days of the week. They can also be irregular events, such as the weather and political or social happenings.
If you are a keen observer, you may even pick up on cross channel patterns. For example, some businesses have high web traffic on weekdays which converts poorly, coupled with direct traffic on the weekends which converts well. This pattern can indicate that people are doing their research on week days and then returning on the weekend when they have more leisure time to make purchases.
Unique significant changes
Generating valuable insights is a bit like looking for a needle in a haystack. Where are there significant and unexpected changes in performance and what caused them? The challenge is to distinguish normal fluctuations from significant changes.
The key is to compare significant sets of data. For example, when you look at conversion-rate performance make sure you have a significant number of users in each segment. Visualising data as graphs can be a powerful tool for identifying these changes. It also helps to review different time segments such as data per day, week, and month. If you compare two data sets, make sure you use a tool to test statistical significance, such as an A/B testing significance calculator.
Focus on the right KPI’s
Understanding the big picture is crucial to drawing the right conclusions. For example, cost related figures, such as Cost-per-Click and Cost-per-Acquisition aren’t just affected by the quality of your copywriting and artwork – they are also highly impacted by competition.And last but not least, be creative, think outside the box and let the numbers speak. Sometimes your website users will behave very differently from how you expect them to behave!
This guest post was written by Lena Swoboda, Digital Strategy Lead at Salmat. This post was originally published on Salmat’s website and has been republished here with permission.