Digital Transformation
Showing results for 
Search instead for 
Did you mean: 

Know your customers 100% better


We have just created a “snap shot” detailing how companies are using big data to “know their customers 100% better”. Think of a snap shot as a mini white paper.

The snap-shot talks about three main ways in which companies are knowing their customers better..

1: Analysis of sentiment
There are a number of things we can do with sentiment analysis..

Forming clusters and monitoring sentiment
Looking at twitter and web sites allows companies to firstly look for clustering. Clustering is where we look for customers talking about a similar subject. Obviously the analysis system has to be a lot smarter than simply looking for keywords. It has to actually understand what humans are talking about. For example, a marketing department may notice that there is a cluster around “cool coats that our competitor is selling”. Or “terrible service with product X”. 

Once clusters are identified, they can be analysed for sentiment. Do people like our products, or not? And how do these sentiments trend over time - is people’s love of our products getting better or worse?  (see the diagram below)


sentiment clustering.png



This clustering of human interaction data is applicable to other use cases too. I was in Copenhagen this week, where we heard about the latest functionality of our service desk. They too use clustering. They use it on self-service searches to identify topics that customers need better information on. They use it on incidents in order to create problems - it’s very difficult for humans to go thru large numbers of incidents looking for clustering (i.e. problem areas).

The police in the UK use clustering of social media too. They are looking for a build-up of common negative sentiment. Once they have identified negative sentiment, they track its intensity. If it builds up, they can work with community leaders to address the issues.

Social media analysis for Social Network Analysis (SNA)
The other thing you can use social media for is social network analysis. If you analyse social media interactions, you’ll find that some people are central - they have lots of connections and they interact a lot (it’s a little more complex than that - but you can imagine a graphical depiction of a user connections graph with well connected people in the centre).  If one of these centrally connected people churns (stops using your mobile service, for example), they may well take those people close to them with them (see the diagram below - the central person has churned, and so her network is more likely to churn too). 


central person.png



Conversely, if you can persuade a centrally connection person to buy your product, their “network” will view that product favourably. So, SNA, or social network analysis, allows us to target anti-churn and product promotions. 

The London Police use SNA too. They are looking for these central, well connected people, for it is these people who can help them defuse potential social unrest. Interestingly, these social media-centrally connected people are almost always not the conventional community leaders. 

2: Analysing customer transactions, quickly
Big data is cool, trendy technology, and so we often forget that the application of big data can be as “simple” as just doing what we’ve always done, but fast enough such that we can take action at the right time. For example, a customer is about to make a purchase and we can help them thru our analysis, this is way better than telling them a week later “we could have helped you last week”. 

Or, as the clothes retailer Guess does, fast big data analysis can give their store managers the insight required to arrange their stores optimally before the customers walk thru the door in the morning. 

The time/value of insight
I think that this topic is fascinating because the value of insight can vary massively. If I we can provide insight while the customer is making a purchase decision, this is hundreds of times better than providing that insight once they’ve left the store. Or, if we can provide insight while the truck is being loaded at the depot, this is hundreds of times better than providing that insight once the truck is on the road. 

The time/value of information is not definitely not linear.  The diagram below shows this (my thanks for HP ES’s big data group for the graph).


time value of insight.png

3: Finding patterns thru customer transactions 
There are a number of things that customer transactions can tell us …

Looking for affinities
Statistically, humans are quite predictable, apparently. If you buy product X, there is a high probability that you will, at some time, buy product Y as well. Or, when you are in area X in a game, there is a good chance you will want to buy virtual weapon Y. This is know as affinity analysis.

The examples above are affinities over a very short time-period. You also get affinities over a longer time period - customer journeys. For example, if I buy a house, I will probably buy paint, a dish-washer, a new TV, and so on, within the next few months. We can use customer transaction analysis to detect the start of a customer journey. In order in infer these longer-run affinities, we need to keep transaction data over tens of years. (see diagram below)



long time affinity.png



The US retail chain Target caused a storm a few years ago. They used customer transaction data to infer when a woman is expecting a baby. Apparently, the woman will change her shopping habits - no alcohol and healthier food choices. Target then knew they a pregnant woman would probably start upon a “customer journey” - buying nappies/diapers, buying a crib, etc. The clever data-based inference was considered a step too far by some of Target’s customers. 

However, retailers love affinity analysis, because, on average, accurate and timely affinity analysis allows them to increase the average transaction value per customer. Or, it allows us to increase the loyalty of customers - i.e. reducing customer churn.

Using transactions for recommendation profiles
We all know about recommendation engines now. Of course, Amazon reigns supreme when it comes to recommendations. Recommendation engines need personal profiles. And it is customer transactions that are used  to build up a view of your personal preferences. 

Want more?
The full “snap shot” can be found here


If you'd like to see more HP big data blog posts and other HP big data news, please go to our big data "" page, here.

Mike Shaw
Director Strategic Marketing

linkedin.gifMike Shaw

0 Kudos
About the Author


Mike has been with HPE for 30 years. Half of that time was in research and development, mainly as an architect. The other 15 years has been spent in product management, product marketing, and now, strategic marketing. .

See posts for dates
See posts for locations
HPE at 2018 Technology Events
Learn about the technology events where Hewlett Packard Enterprise will have a presence in 2018.
Read more
See posts for dates
HPE Webinars - 2018
Find out about this year's live broadcasts and on-demand webinars.
Read more
View all