FOLLOW

FOLLOW

SHARE

A Different Face For Every Customer: Big Data For Website Personalization

Customers today expect websites to cater for them and data makes it possible

26Sep

Marketing relies on target group segmentation to provide tailored content. While traditional segmentation techniques using demographic and sociographic filters are too broad to yield satisfactory results, going too in depth can narrow the target market too much to remain relevant. In this case, what should a company do?

The personalization spectrum

As Deloitte explains on their Perspectives blog, just because you can, doesn’t mean you should go up to a granular level of customization. It is important to see the personalization option as a continuum ranging from mass market to personal selling. Each company should create a business model including the decision of how deep it should go with personalization.

This is important since acquiring, cleaning, analyzing and most of all acting based on results is a costly process, and the money invested in such an approach should generate ROI.

History of personalization

Offer customization began long ago offline, by identifying the most likely product associations having in mind the typical customer. Just think about toothpaste and toothbrush packages. Online commerce took this concept a step further and investigated less obvious associations that generated cross-selling. Amazon pioneered their trademark system of collaborative filtering that created a performant recommendation engine.

As described by a recent article by Iflexion, recommendations are most useful for new clients, an underserved group. But how can you recommend a great and appropriate product to a stranger? That is where the context provided by Big Data comes into play and boosts sales and loyalty.

Hyper-personalization

At the current stage, Amazon and Netflix are doing what is called hyper-personalization. They are using a vast amount of data to create detailed user profiles and generate recommendations based on these. In fact, this has become their business model. Also called the long-tail model, it relies on selling a small number of a diverse range of products. By targeting the user very precisely, these websites make sure at least one of the offers will translate into a sale. If this does not sound like your business, don’t abandon the idea altogether, just decide on what degree of personalization is necessary for your organization.

A strategy to implement Big Data

Even if most companies don’t need the degree of personalization employed by Amazon, there are a few lessons to be learned about implementing a successful Big Data project on your website.

First, start by defining your areas of interest and subsequent problems. Do you want to find out why users are dropping along the sales funnel? Do you want to extend to new geographical markets or age segments? Are you suspecting some of your products are underperforming and should be eliminated from the offer? Do you want to make a proper allocation of the marketing budget? Do you look for acquisition, conversion or remarketing?

Next, define a hypothesis for your problem and the data that will help you get a straight answer. Measure a baseline and start making one change at a time, per user group. Record results and keep track of successful combinations. Don’t make changes based on your assumptions about the target population, but on data. Refine and re-test until you are satisfied with the result.

Data required to create unique experiences

For these kind of projects to be successful, they need loads of data to be fed into the algorithms and used as raw material for learning the client’s behavioral patterns, as well as correlating those with determining factors such as age, residency or incomes. Some of the data can be collected on site by the client’s interaction with the website, while other, generally named context can come from other sources such as the browser’s cookies, connected social media profiles or linked credit cards.

On-site

The interaction with the website is significant for understanding the brand’s success rate. All companies should check at least the statistics provided by Google Analytics. The first indicator is related to the way clients reach their websites. For example, if large numbers are coming from social media, that is a good indicator of where the marketing budget should go. Also, it is important to match the profile of the ideal client with the demographic data of actual customers. Continuing along the sales funnel, identifying drop-off pages helps a business understand what should be done differently. Correlate the drop-offs with demographics to see if you should split packages by different factors like age, sex or location.

The user’s profile, if available, is a goldmine of information related to interests, price range or accessories and related objects to past purchases.

Off-site

Every user has a story. This is, in fact, the context of their visit. Not all people who land on a site are clients or leads, so an organization should learn to distinguish between real customers and just passers-by. Background information like geolocation, visits to competitor’s websites, questions asked on forums and social media, all give clues about the person. The personalization should target just those that are possible clients and keep an appropriate basic level for all others, like recommending best sellers or latest products.

Types of personalization

Although there is no magical recipe for increasing conversions by personalization, the most efficient changes include:

  • relevant text and images, above the fold or as soon as possible;
  • presenting relevant products to the user;
  • targeted pop-ups and offers;
  • surveys and quizzes to improve the experience;
  • customized info bars (based on geolocation, for example) and calls to action (based on previous searches);
  • a mobile-first approach.

Keep in mind that personalization items should not detract the user from pursuing their primary goal of being on the website but come as a nice addition or help.

Beyond website personalization

The latest enhancement in website personalization originates from the brick and mortar world. Big Data is not only available online, but, in the case of a business that also operates offline, the rise of IoT opens gives the opportunity of collecting more customer data and converting it into context.

Comments

comments powered byDisqus
Turkeyss

Read next:

What’s On Data Analysts’ Plates This Thanksgiving

i