Friday, May 14, 2010

IBM unveils analytics that understands social media slang

IBM has unveiled a new business analytics tool that allows organisations to derive useful customer data from the slang and emoticons used by consumers on social networks.

The SPSS Modeler text mining and analytics software aims to enable organisations to scour the internet for feedback on products and services, uncovering and analysing information from social media sources, such as social networks and blogs, to provide a more accurate view of what consumers are saying.
Recognising that each industry has unique priorities and its own vernacular, the new software analyses trends and captures insights from industry-specific terminology.
Within these domains, the software includes new semantic networks with 180 vertical taxonomies (from life sciences to banking and insurance, and consumer electronics), and more than 400,000 terms, including 100,000 synonyms and thousands of brands. This allows customers to draw better links and understanding between sentiment and products without having to spend time building their own definitions.

Character-based Segmentation, by Ian Stockley

Imagine you are an online fashion retailer. Two of your customers both shop with you on a regular basis, both buy the same pair of shoes and the same dress online. From this information you can safely assume that these two customers will also share the same tastes when it comes to coats, bags and jumpers and again will buy these products through your website …right? Wrong.

Simplistic as this example may be, it highlights the problem with the way many retailers currently segment their customer data and subsequently target their communications. So what can be done to overcome these problems? The answer might very well lie with a fresh new method called character-based segmentation.

Typical segmentation works by either segmenting customers’ behaviour by using simple RFM variables (Recency, Frequency & Monetary value), which effectively shows your best and worst customers. Or at the next level, using FRAC variables (Frequency, Recency, Amount and Category) to then separate your best customers over the various products you sell.

The obvious problem with this approach is how do you differentiate between your ‘best’ customers or indeed, your ‘worst’? Do you just assume all your ‘best’ customers share the same characteristics, tastes, attitudes, preferences and market to them in the same way or do you find a more sophisticated segmentation process? One which is capable of identifying the distinct segment groups within your ‘best’ customer base?

And it’s here that character-based segmentation really comes into its own.

Character-based segmentation: an introduction

This emerging method puts a great deal more variables into the pot than traditional segmentation processes. In addition to transactional data, which tells you what people are doing, character-based segmentation also incorporates two other important and powerful sources of information. Firstly, unlike traditional techniques, it draws on enhanced personal and derived consumer lifestyle data to add that extra level of detail about how individuals behave and engage with brands. Such details might for example include the consumer’s previous purchases of multiple brands, their reading habits, musical tastes or their preferred media channels for engaging with individual brands. The third and final source is external digital data, which includes sources such as web browsing data andemail engagement. These additional sources of information provide vital information about what consumers are thinking, their personalities and their preferences.

When you bring all these different sources of data together and analyse the results, distinct and highly detailed customer segments soon define themselves. In this way you get multiple groups of your best customers and multiple groups of worst customers. But importantly, each group has a unique personality and characteristics, based not just on transactional data but individual attitudes and preferences. Understanding what your customers look like and how they think outside of the context of the brand, provides an enormously effective weapon with which to shape your marketing communications strategy.

The character-based segmentation maximises engagement and forges strong connections with both customers and prospects, allowing the most compelling creative ideas to be developed. With this detailed picture of your customer base, key strategic or creative insight is revealed that otherwise is very likely to be missed. By identifying these key differentiations segmentation adds so much more value to the bottom line, often as much through the creative variations used as in the increased ability to add further discrimination in the data selection.

It becomes possible to tailor all aspects of your marketing to the specific demands of each group, from the communication channel and message to frequency and offers. Importantly, each segment is appended to named individuals, moving from marketing based on assumption to marketing based on facts. The effectiveness of your communications is dramatically improved with increased response and conversion rates, reduced costs and long lasting connections with your customers made.

More than online retailers

The detailed insight gained from character-based segmentation can be transferred across the whole organisation, from the marketing team to sales, buying and customer service. As a retailer if your buying team knows there’s a distinct segment of your customer base that has a high propensity to buy specific designer brands on a regular basis, then your purchasing strategy can be adjusted to include more products that are likely to be of interest to this group of customers. The whole organisation moves from a product to customer focused orientation. Character-based segments can be brought to life for a whole company by using and sharing visualised, engaging descriptions of the character segments across departments through use of intranets and interactive media. This provides the collateral to deliver a customer driven focus for a brand in a way that is so often talked about but so rarely delivered by companies.

Character-based segmentation is by no means limited to online retailers. It can be equally applicable to any brand or market sector. If we took the travel sector a customer that goes surfing for a week in Spain and then follows it up with a two week winter holiday in Egypt could be placed in two segments by typical segmentation methods: 'Budget travel in Europe' or 'Adventure travel further afield'. While this method might be able to identify destination preferences it will not explain the underlying reason why the customer is travelling.

Character-based segmentation would understand the personality who is after adventure in holidays, whether it be near or far afield. A tailored marketing message selling a selection of holidays to this adventure seeking individual would be very different to the message that would be created on the back of knowing only that they had spent a certain amount on holidays in both Spain and Egypt.

In today’s environment the most mutually beneficial brand relationships are built on detailed customer understanding. Success comes to those companies that take the time to understand the needs, preferences and life stage of their customer and treat them as individuals. It’s no longer enough to know what your customers want; you now need to be able to understand why they want it.

Newer online brands such as ASOS are moving this way and households brands, particularly those who have seen such success with loyalty cards will be following close behind. Character-based segmentation gets to the very heart of the customer decision making process, uncovering key drivers of behaviour to allow brands to make real connections with their customers. In the complexity of today’s multi-channel world, such a step-change in approach to customer segmentation has never been more timel