In a modern multi-channel environment, it can be challenging to know who’s doing what and why. In many cases your customers will experience multiple touchpoints before making their purchase, but can you identify what affects their choices?
The answer is ‘yes, you can’ – multi-channel attribution is becoming a high priority for many ecommerce organizations as we move away from cookies. You can still use data and analytics to discover where your customers came from and the journey they have taken.
What is multi-channel attribution?
You can find out which of your marketing channels are most likely to lead to sales. This will help you to work out where you can achieve the best return for your marketing investment.
A typical customer journey might include discovery on one of your social media channels which leads them to your ecommerce website. Although they choose a product and go to checkout, they don’t complete their purchase. They might then receive an automated reminder email and return to make their payment. In this example the sale could be attributed to just one of these touchpoints rather than each of the steps your customer has taken, which would be single-touch attribution.
In contrast, multi-touch attribution gives more context by showing you each of the steps involved in completing the sale, helping you to understand their relative importance.
Why use multi-channel attribution?
You can learn not only where customers came from but also which channels convert to sales most effectively.
You might think the answer is obvious but, for example, many ecommerce companies use social media advertising to attract new customers and Google Ads to capture brand and product-specific searches. Google Ads will provide a good return on advertising spend (ROAS) because it meets the needs of people who already know what they’re looking for and are more likely to buy. However, without social media channels to raise awareness searches leading to Google Ads might never take place.
Multi-channel attribution also allows you to see and understand the role of each touchpoint so you can make informed decisions about your marketing. For example, if you know that half of your customers make their purchases using smartphones and mobile devices it makes sense to optimize your ecommerce site to give them the best possible experience.
Channel attribution models
First-click – the simplest approach attributes a sale to the channel that generated your customer’s first click.
Last-click – alternatively you could attribute the sale to the last channel your customer used before their purchase.
Multi-touch – a more accurate approach will show you that your customer discovered you on Facebook, read your blog, and made a purchase after visiting a product page.
In its simplest form, your customer might see a social media advertisement and go to a subscription page on your website where they can register. They might then receive an automated email with a special offer for their first purchase, which leads to a sale. You could attribute one third of the sale to each of these channels – this would be a linear attribution giving equal weight to each touchpoint.
Time-delay – a time-delay model reduces the importance attributed to a channel depending on how the time between using a channel and making a purchase.
Position-based – this approach gives added weight to your customer’s first and last clicks before making a purchase, but also gives some credit to channels they might have used in between.
Multi-channel attribution challenges
Every business is different, so it will be important to spend some time exploring your data to understand what it’s telling you. If it isn’t clear you can create some theories and plan tests to find out more. However, there are also some issues to consider.
Mis-attribution – you could see an increase in your Google Ads results after increasing your social media advertising. If you look closely, you might discover that the results are for exactly the terms you have used on social media rather than an increase in your overall organic search results.
Data privacy – with growing concerns about online privacy and cybercrime, there are now more limits to the information you can capture online about behavior before someone becomes your customer. You might therefore need to collect data from multiple sources to analyze within your enterprise resource planning system to create useful insights.
Offline touchpoints – devising meaningful and helpful approaches that allow you to capture information about offline touchpoints will be important to give you a complete picture of your customers’ behavior. This might include customer surveys and in-store apps, for example.
Using multi-channel attribution
Your first step will be to decide which type of attribution model best suits your business. You will need reliable and accurate approaches to capture the data you need and a clear understanding of what this information can tell you.
A good place to start could be an evaluation of data you already have. You can use this to test your existing impression of customer journeys and to identify gaps.
You can then explore your theories with customer surveys or focus groups. For example, if you discover that most of your customers find your products on social media channels but buy from your ecommerce website then a last-click model won’t help you to understand their experience.
If you are planning to use Google’s replacement for Universal Analytics (UA), GA4, you will want to be sure you have a clear view of the attribution model you intend to use.
Once you have collected a reasonable amount of information you can start to test, review and refine your theories based on real data and monitor the impact of specific marketing activities.
Work with a trusted partner
At Williams Commerce we use a ‘test and learn’ approach. We will help you to implement changes and record user behavior over time. Finally, we will investigate the statistical significance of your results, both quantitatively and qualitatively.
