An Approach for Personalization - Digital Strategy Execution Part 2

In part 1 of this series, we established baseline understanding for Personalization. Let's talk about you can approach its implementation.

At its heart, 'Personalization' is a simple concept i.e. the more you know about the person interacting with you, the better you can draw inferences on, and serve their needs.  

As such, the data you collect on your customer will merely represent the focus of your search. Every interaction a customer has with your business, on any channel generates new data and potentially new insights to factor into your assessment.

However this does not mean you have to“measure everything” – that is never an answer.

Quite often people will look at personalization as trying to arrive at a magic formula - collecting a maximum number of data points and 'facts' about the customer, with which they can use to market with 100% assurance, when the truth is far more nuanced than that.

You have to ask the right goal-based and KPI linked questions to collect the right data. It's also important to be aware of the the data telling you something you weren’t even asking.

To ensure that you are collecting right information you really need to understand real world market, relevant demographics and the industry position of your own business.

Consider the following scenario:

A UK based electronics retailer realizes that a lot of people in London are clicking on the cameras sections of their website. They thus decide to buy relevant keywords and invest in cross-platform advertising to drive further traffic in hopes of increasing sales.

They however , don't see any increased conversion – After deeper analysis, they realize that those that DO convert opt for camera rentals instead of outright purchase.

Camera rentals are not prominently featured as a part of their services - and thus a lot of traffic tapers off before conversion.

This is a typical example whereby the above is not a failure but an opportunity to fine tune and hone their strategy.

In the example above, there are few things the firm can do / could have done

1. Get to know better how their business stands in the market – truly understand who may be interacting with their brands and what their needs are. Perhaps their biggest consumers are tourists that need rentals.
2.  Allow for flexibility in their data model and collection strategies to take in new parameters and pivot on new facets and insights
3. Ensure experience channels can adapt to the ongoing discovery of new – if their existing channels are unable to effectively promote Camera Rentals without significant investments, then it’s time for an overhaul
4. Don’t lose sight of goals. Set clear unambiguous top KPIs and metrics as a part of Digital Strategy – The KPIs will drive your search for insights. If in the above scenario, the KPI was camera purchases instead of overall conversion, then the team may not have discovered the link with increased camera rentals.

Note: The actual execution of personalized experience depends on the channels – e.g. from a teaser on a website, to exclusive offers delivered via print media, to customized service at the next shop-visit – but they all depend on how well you know your customer at that point in time and how mature your cross-channel capability for personalization is.

This brings us to strategies of getting to know your customers that much more.

Knowing your customers – Major Strategies

Simple Point Based Solutions

Useful for internet connected digital properties including emails, smart devices and of course websites.

These are simple point based solution and are effective for what they do. They give you a point in time, self-contained set of information. Good for building a very high level profile of your customers and something to potentially start with

1. Geo-location
2. Referrer Address
3. Keywords used on the website
4. Ads clicked through
5. Pathing on your digital property
6. Form / Application drop outs

Note that there will be limitations based on what device or browser version the customer is using or even the permissions they have provided your applications ( e.g. allowance of cookies or 3rd party JS ).

Profiling, Segmenting and Targeting

This is doing all of the above, but as a part of a larger strategy to maintain a “bucket” of information about each customer. Each campaign and interaction enriching the context & adding to that bucket – thus over time getting more detailed understanding of the customer behavior.

This is typically the starting point for a wider enterprise digital strategy. Each customer is assigned a unique ID – a tracking ID.   with a varying amounts of traits and attributes that are enriched with each interaction.

At a high level, unique ID is generated w.r.t. to certain unique attributes for each browsing session – typically the Machine ID / Mac ID if available for anonymous customers or CRM ID equivalent for known logged in customers. Each browsing session or interaction would add context to these IDs and thus can be used for segmentation (grouping according to traits ) and subsequent targeting.

The trick, as mentioned in explicit personalization in part 1, is how can an organization match the anonymous IDs with named customer IDs such that they don’t hit false positives.

Note: In Adobe Marketing Cloud implementations, one of the IDs you will hear all the time is Visitor ID or Marketing Cloud ID ( mid ).

Organizations have to also ensure that customer data is secure to appropriate standards, especially since marketing IDs or something similar are generally stored on the provider’s cloud ( in Adobe’s case – the Adobe Marketing Cloud ).

It's possible to treat anonymous personalization as completely different track to personalization in the authenticated space, however it will always be lacking. Real value will be unlocked when there is a strategy to marry the two.

This question, then brings us to a foundational aspect of Personalization – How good is your data strategy. What are some of the considerations an enterprise needs to take when setting a data architecture ?

That will be a matter of the next blog post in a few days ! - So stay tuned, as they say!!

What are some of your major challenges ? What would you like to hear about more ( or less ! ) ? Let me know what you think in the comments below -