Online personalisation is currently the hot topic in digital. In order to achieve success in this area, it is important to develop a data strategy that underpins your engagement plans. Here we take a look at some of the key considerations in the planning of automated marketing.
As an agency, we are both Sitecore Gold Partners and Episerver Premium Partners. We use these platforms to deliver personalised customer experiences that drive results, engagement and ultimately online success.
Personalisation should be customer-centric not device-centric; you need to be able to track a user across their touchpoints. Over 60% of transactions occur after a user journey across platforms, a typical example being mobile browsing to desktop purchase.
For many businesses, consumer interactions exist offline as well – whether it’s call centres or retail stores, consumer engagement in these channels also needs to be considered. The marketer’s goal is to pull them all together.
Hence identity graphs housing customer profiles form the foundation of ID resolution. A clear singular mechanism should be implemented to allow customers to be uniquely identified. This component sits above web, apps, CRM (Customer Relationship Management), EPOS (Electronic Point Of Sale) and so on, ensuring data can be linked and attributed to the individual. We use IdentityServer, but there are many emerging solutions.
The data needed to identify the user has traditionally been an email address but other identifiers can also be considered. This might be a mobile number, fingerprint, Android and iOS advertising codes or customer account numbers. Social validation is another method growing in popularity.
When you know that the individual is one and the same, your engagement strategy becomes universal and many of the risk factors of in-contextual messaging can be removed.
Identifying the target market
We see a lot of ‘campaign-based’ personalisation which just pushes marketing messages along the customer journey. Anyone who gets followed around online by remarketing ads knows how frustrating and random these can be.
Whilst there may be a small place for personalised landing pages and CTAs, it won’t build advocacy or trust in the brand. It merely makes things easier to find.
If personalisation is about creating a unique customer experience, then understanding those customers is fundamental to the vision – hence the data strategy must relate to a knowledge of the target markets. It then dictates not only what data we can collect, but what data will allow us to identify this group. If you can say with certainty that your visitor is group A, B, or C, then you can align engagement and content strategy specifically to their needs.
At Redweb we use Mosaic as the initial benchmarks. This means clients don’t have to create groupings based on products or interests. This work is often common sense and if you approach building an online ‘automated’ relationship the way you would an offline one, then things often become clearer.
All brands want a lasting relationship with their customers. To help retain me as a customer, brands such as travel agents or sports companies should know I’m outdoorsy and adventurous. If they speak to me about my passions – rather than just reiterating our initial topic of skiing, in which I have only periodic interest – then I’m more likely to continue engaging with that brand.
So the principle is personification rather than personalisation. AI (Artificial Intelligence) might eventually deliver an individual experience per person but in respect of strategy and planning, it is still necessary to plan groupings or personas around motivation, purpose and demographics.
Whilst deterministic data provided by customers might be necessary to exactly match an individual against a target market, it is often further in the engagement cycle that this definitive match can be achieved. Hence marketers can look to a longer shadow of data patterns to predict a user’s intention.
Social logins, postcode, device usage, or page views might be sufficient to flavour communications in the style of target market. This probabilistic data holds the clues!
Another component within data strategy defines what datasets your users are more likely to offer – and this depends a lot on the brand’s authority. For example, alcohol brands need to ask if you are over 18, but maybe not your exact date of birth.
This works both ways. A US report stated that 80% of people would give up personal details in return for some payback – whether that’s personalised content or recommended offers. As such, when a brand knows its authority, it can generally ask for more data. So if you’re building a strong relationship with your customers around your expert subject, then a better knowledge of your consumers’ relation to that subject is essential.
The shyness of data capture can be seen in email newsletters sign-up – asking for nothing more than an email address just confirms everyone is going to get the same broadcast. The result isn’t beneficial for either party unless your KPI is just emails in a list or you are so niche that your customers self-select.
At Redweb we grade data against the brand’s authority to receive, the consumer’s resistance to give, and the data’s overall usefulness.
Single customer view
Much is spoken about the creation of a single customer view – a data repository that holds profile, behavioural, transactional and communication interactions. With a single view, clients have greater control over the delivered experience – and many software vendors such as Sitecore see a role in providing experience databases.
However, big data needs management – and the emergence of AI and more sophisticated business intelligence tools will allow for greater and faster insights. Therefore, any data strategy should be identifying existing data silos and recommending migration or integration into a central location.
The new EU privacy rules should be welcomed as they _should_ make businesses think and be smarter with the data they hold on us. In respect of web then the two main topics will be security and disclosure – a data strategy needs to define both.
The terms ‘Forget me’ and ‘Notify me’ highlight what website owners need to be considering. If the ability for a user to remove data collected for non-transactional use is planned in from the start, it not only allows compliance but also engenders greater trust. ‘Forget me’ can be as simple as displaying the data collected as an extension to a profile. Being able return the field against the user to ‘not known’ removes the association.
Big data should be aligned to hashed ID keys and not specific identifiers. This allows associations to be better controlled and managed.
The data strategy acts as a foundation to the personalisation planning. That includes scoping:
- How the data is collected
- The design of the tools/functions required
- The hooks necessary to allow probabilistic clues (which will influence the IA and copywriting)
- The technical infrastructure.
However, with a clearer understanding of the data available, touchpoint definition and a knowledge of how target markets respond, we have a greater chance of achieving success.