5 steps to personalising customer experiences by connecting online and offline data
Merging online and offline data is the key to marketing success in 2018. But particularly in the multi-channel, multi-device, post-GDPR age, doing it effectively can be tricky. How can you track an individual customer across different channels and devices? How can you understand the full customer journey, whether in store or online? And how can online data improve the relevancy and cost-effectiveness of your marketing at each stage of your customer’s buying cycle?
Overcoming these challenges can reveal business insights that can help to propel your sales and marketing performance. And, thankfully, it’s not as complicated - or as daunting - as many fear.
In just five steps, you can gather and merge the data you need to improve the relevancy and cost-effectiveness of your communications.
1. Collect the data
Collecting data is crucial, but it must be done correctly: you need to be able to access the raw data and see log-level information on individual consumers. There are some solutions that help you do that, but Google Analytics is not one of them. Its ID function will get you close, but it isn’t sophisticated enough to get you through the next four steps.
So, what do you need when you collect data? First, you’ve got to be able to collect data on a device level and to understand how that works. At this stage, don’t worry about cross-device data collection (we’ll come to cross-device data matching later in the process).
Next, you need to look for a solution that does two things: earmark a device with a unique ID and set a visit ID so that you can understand when fresh visits occur from the same user. Whether your user is on mobile, desktop or tablet doesn’t really matter at this stage – just collect a device ID and visit ID every time someone lands on your website.
You also need to be able to collect and post multiple IDs against a visit ID, so that you can see the user has clicked button X, visited page Y, etc.
2. Organise your IDs/keys
There are hundreds of different IDs on any website. The number and type are specific to your business, set-up, stack and configuration. Some of the most common include:
Exposed user ID
However, the email address is personally-identifiable information, so you must process it first with a SHA-256 one-way hash. This uses a mathematic algorithm to prevent the email address from ever being identified and is impossible to decode. You’ll simply be left with a long string of numbers and letters instead of an identifiable email address.
From these three steps alone, you end up with an exposed user ID, an email hash, and eCRM ID from just one visit ID.
3. Wrangle your data
Wrangling the data, and getting it into a place where you can use and query it, is the stage where you might need to call on your IT department or outside consultants. You will need someone competent in SQL or a similar language. It’s quite a technical job, but it’s a very important step.
Once you get your data into a big data environment, it opens up a huge amount of value. But to do this, you will need:
- A lot of server space;
- A reasonable amount of power (your average desktop won’t be sufficient); and
- To be prepared for unperfect data
This is a labour-intensive job and it’s specific to every business. Be wary of any tools that claim to do this stage easily. They don’t exist.
4. Make the connections
At this stage, we’re starting to link the digital data to the offline data. We’ve collected the data on a device level, added other IDs such as email hashes, and can now start to run matching logic inside the data environment and merge the records with matching IDs.
For example, when there are multiple visit IDs sharing the same email hash, you can assume that’s the same person logged in on multiple devices and start to merge those records. Once you’ve competed your cross-device logic, you can also bring your offline data into the same process.
This stage will involve taking a deep dive into joining logic and considering exactly how you can best ensure you reach the granular level of detail your business needs. But, once it is complete, you can then look at individuals against their previous marketing exposures. What campaigns were they exposed to, and what was the lifetime value of that exposure?
5. Analyse the data
This is the fun bit. Once you’ve created a database with matching data, you can start to get down to detailed analysis. You can:
- Analyse marketing spend over time, holistically, or on segments;
- Produce advanced data-driven attribution encompassing lifetime value;
- Model next-best actions for customers;
- Create predictive models based on uplift of channel spend. (For example, are there certain channels or keywords that create customers that stay for three or more years? What’s creating the real long-term value over customers you see once and never again?)
And because your data is set against multiple IDs, you now have the ability to use those IDs in the execution platforms you use, ie, email hashes in Google Customer Match or Facebook Custom segments. It allows you to create an execution mark as well as an analysis mark.
When these five steps are complete, you are well placed to improve the relevancy and cost-effectiveness of all your communications and media. You can:
- Target new customers with the most effective communications, as proven by the value they generate;
- Track and influence customers in an extended consideration phase;
- Pre-empt customer needs;
- Reward loyal customers;
- Stop lapsing and basket abandonment before it happens;
- Adopt proactive retention strategies to lock-in customers.
For more information on data-driven marketing, read our latest blogs.