Uncategorized

How Did I Start My Career in Analytics?

 

asking-questions

People often ask me, “How did I start my career in Analytics?”. It was actually, quite by accident. In my first year at university I wanted to do the key science subjects such as Mathematics, Physics, & Chemistry but I needed a fourth subject and based on my time table options, Statistics was the only science course available for me. I had no idea what Statistics was about but I fully believed in myself that I could do it, even though there were many stories that Statistics is not easy to pass and the pass rate was only 40%.

I quite enjoyed the Statistics lectures and my professor, Prof Calitz told interesting stories and often challenged us with tricky problem solving questions. I worked hard at solving these problems and often went to my professor to share my solution and ask questions.

Prof Calitz never gave me a straight answer, he always rather probed me with one question after the next and sent me back to complete my solution. At that time, I thought Prof Calitz was a hard man and not very helpful but looking back, I can say that it is through him that I now have a good knack of asking lots of questions while problem solving. And you know what, during my university studies I had to work hard at my other courses such as Mathematics but didn’t really open a book for Statistics and still passed it with a Distinction!

I am today, very passionate about Analytics and enjoy working through business problems, which are quite challenging and more complex.I am very thankful to Prof Calitz for his good teaching approach!

5 Tips for Building a Powerful Dynamic Customer Analytics System

 

  1. Ask the right questions to achieve your business objective:Begin with the end in mind when asking questions, gain clarity and focus on what you need to know to achieve your objectives. If our goal is to increase our product sales to achieve a financial return of 80% say, To gain clarity we need to ask ourselves the typical 5 questions: Who?, What?, Where, How? Why. For example, who is buying our products, what products are they buying from us, where are they buying (which geography)? How are they buying our products (which channels are they using)? Why did they buy our products?Brainstorm with “what if” scenarios. Understand what the best case could be in terms of product sales, what could be the most likely case and what could be the worst case.

    Companies across industries are trying to understand and connect to the consumer at a more personalized level.There has been a strong shift from the business-to-business (B2B) model to the business-to-consumer (B2C) model.

2. Track the Metrics crucial for your organisation – Use industry KPIs

A good metric/KPI is:

  • A number that drives change you are looking for
  • A good metric is comparative
  • A good metric is understandable
  • A good metric is a ratio or rate
  • Quantitative
  • Actionable
  • Leading Metrics
  • Causal Metric

Below are some examples of good KPIs:

  1. Sales Growth 
  2. Leads 
  3. Customer Lifetime Value (CLV) 
  4. Cost of Customer Acquisition (COCA) 
  5. Website Traffic to Website Lead Ratio

 

3.Count People and number of unique Visitors instead of:

  • Number of Hits
  • Number of Page Views
  • Number of Visits.
  • Number of Downloads

 

4. Ensure that your Dynamic System is Data Driven:

To develop a Dynamic Data System, you need the right ‘People’, Processes’ and ‘Technology’.

The key people for a dynamic system is firstly, a Domain Knowledge person – understand the business and data available for answering the business questions,

Secondly, you need a Statistician who understands the statistical techniques that need to be applied to the data to discover insights and patterns that help to solve the business problems. The statistician is also able to provide data visualisation fast using drag and drop type tools for visualisation (Tableau, QlikView, SAS or R) to help the business user to make valuable decisions.

The third person you need is an Information Technology Specialist who ensures that the data definitions are consistent and that all the data is in one place and easy to retrieve for analysis in real-time thereby providing and effective enterprise dynamic system.

For Processes, we need to ensure that we have the right processes in place and that they are continuously improved, which means, we need to track process time and ensure that they are reasonable or even aim to shorten process times.We need to track customer complaints to better understand which processes require improvements and remove the processes that are not necessary.

5. To get the Dynamic System Right – Ask Lots of Questions

Starting with a clear objective is essential and then we still need to ask lots of questions once we have identified the objective. Question such as,  ‘What do you want to Achieve?’, ‘New Customers?’ or ‘Increased Customer Loyalty’ or ‘Increased Revenue?’. Then for example, how will you get the increased revenue? Will you run a promotion? How much increased sales do you expect from the promotion? Which customers are likely to buy from that promotion? When will these customers buy? How will they buy and which stores will they buy from? All these questions can be answered using statistical modelling and classification techniques that can be linked from one to another and in real-time the promotional offer is made to the right customer at the right time with the right product.

 

 

 

Dynamic Analytics: What is Dynamic Analytics?

What is Dynamic Analytics? To answer this question, let’s take one step back and first ask ourselves, what is Analytics?

Analytics in a nutshell is about solving a business problem through asking yourself lots of questions that will allow you to better understand the business problem at hand.

Once the business problem is understood, it is for the analyst to work with the domain expert in gathering the relevant data and then to summarise the data numerically and visually to better understand the data at the aggregate level. Further, many hypotheses will be tested through comparing different groups and thereby gaining insights as to where statistical differences exist between various groups, customers and products.

Some correlation analysis tests between different variables will be made to understand the statistically significant relationships between the many business variables.In most cases, predictive analytics will be performed to allow businesses to plan their resources and strategy based on the predictions of whats likely to happen in the future. And the final step will be for the business to take the predictions made and to look at their resources to determine what is the best that the business can do to enable the prediction to take place.

Dynamic analytics is learning what customers want faster. Advances in Technology allow businesses to learn what customers want faster. Businesses also need to be aware that not all customers are the same. Customers differ on their demographics, their lifestyles, and their buying behaviour. Different groups of customers are interested in different products and services. Businesses therefore need to gather lots of data about their customers in almost real time to better understand when customers need, want and buy their products and services.

It is only with analysing customer data, can businesses better understand what customers buy, how often they buy, how much they buy, how they buy, what they buy together and when they buy. By using customer’s click stream on the web businesses can learn faster, almost in real time what customers want.

To learn what customers want faster, businesses need a social strategy, local strategy and a mobile strategy. Why do I say this? Well in today’s time, everyone is almost always on their mobile, so businesses need to analyse customer’s mobile usage and understand when are customers using their apps, how are customers using their apps, etc.

Research has also shown that customer’s research the products and services they are interested in, on their mobile while travelling on the bus, train, car and then share their knowledge with their friends, family, or colleagues through social media. Therefore businesses should have technology designed to listen to their customers when they interact socially and will then learn very quickly what customers want.

Lastly, research has also shown that most customers use their mobile to find products and services close by (local). Therefore, businesses should make sure that all their details, location, open and closing hours, information about their company, products and services are easily found when customer’s search for the nearest….restaurant, bank, petrol station, etc

Dynamic Analytics is ensuring that the business provides their customers with the right product at the right time, at the right place, with the right price & channel. Dynamic Analytics allows businesses to delight their customers by getting the next best offer right, that is , “offering your customer your product” before your customer even knows they need it!

Singapore – An Analytics Hub?

Singapore ranks amongst the most advanced & competitive IT services markets with high levels of technology adoption. Adoption of newer technologies ranked as very important or important among 73% of Singaporean respondents. More than 95% were looking to invest in newer technologies in the next 12 to 24 months. More than 75% of respondents are looking to have customer intelligence, predictive analytics & sentiment analysis in place in the next 12-24 months.

More than three million people in Singapore use Facebook and more than 900 000 use Twitter. This gives businesses a great opportunity to connect with & better understand their consumer base at a low cost.

Business Analytics has been identified as an important growth sector for Singapore.There are currently more than 100 apps developed using government data by the private sector & community groups. These range from car park availability to clean public toilets. Singapore moved forward with open data initiatives (data.gov.sg), a one stop data portal with more than 8600 data sets from more than 60 public agencies. Singapore provides data sets for crowd source analytics solutions, resulting in rapid prototyping, piloting & developing proof of concept.

Is Singapore an Analytic Hub? Most certainly. Analytics goes hand in hand with high technology. With high technology, data is everywhere and needs to be analysed  in almost real time…many apps are being developed each day. Many hand held technologies that provide business insights in real time are becoming the order of the day.

Because of the reams of high speed data being generated from technology apps, decisions can only be made through measuring and analysing this data. It is absolutley essential to use this data in business decisions and for business growth.

There is a shortage of business analytics talent, to narrow the gap, software vendors are collaborating with academic institutions to develop curriculum that reflects the mix of technical & problem solving skills that is necessary to prepare students for business analytics careers. If you need more information on business analytics and the topics covered in the different business courses, do not hesitate to contact me.

Transforming Healthcare using Big Data

What if nurses & doctors could remotely monitor their patients through real time notifications based on pre-set thresholds set by the doctor, based on the patient’s condition?

Many doctors and nurses think that Big Data discussions and strategies means more work for them, not less, and that Big Data implementations will take away time from what they see as their key responsibilities such as consulting with patients and providing quality care. The key challenge to using healthcare data smartly is that, ‘Big Data’ brings with it as top of mind….data costs, risks, liabilities & patient privacy, and this thought just scares the key stakeholders as healthcare data is not seen as a source of value, but of additional work.

Another key challenge is that different users imagine data in very different ways. Understanding this key facts about data helps to understand why so-called “big data” solutions are so difficult to implement in practice.The biggest challenge for the use of “big data” in healthcare organisations is not technical. The challenge is figuring out how healthcare professionals such as doctors and nurses, management and technical, will actually use the data in practice.

There is a gap in the understanding of the value of Big Data and how Big Data can help doctors and nurses and free them up from many of the duties and roles that they are currently finding difficult to do, only because, they have so many more patients to take care of in a day, than a few years ago. Using Big Data doctors will be better able to target and understand high risk patients by utilizing patient key biomedical data and natural language processing to extract key elements from unstructured History and Physical, Discharge Summaries, and Consult Notes.

With the silver age, hospitals and clinics are busier than ever. Many times treatments even have to be delayed because of competition for beds or doctor time. The outcome from understanding the time management pain points of doctors and nurses – is by using analytics, we can monitor patient thresholds using a decision support system at the ward reception desk. Real time patient risk models can also help to predict and identify which patients are more at risk and need more effective management to prevent them from getting worse. This way monitoring of patients more efficiently can be done as doctors can have real time updates on how their patients are doing and may reduce their time with patients in control and have more time to focus on high risk patients and help prevent them from getting a disease or at least reducing their their risk.

Healthcare analytics

I would like to have the opportunity to analyse real patient data, to identify patient risk factors for a particular disease. Do you know any healthcare organisation that would like help in this area?

Making the Most of Big Data Now…

When it comes to actually mining Big Data for insights, many companies don’t know where to start or focus on the wrong things and get bogged down…. I say with confidence, Data Visualization is your Key, to making the most of Big Data!

Condensing piles of data to just a few charts is a balancing act of art and science. The visualization should narrate what the next short-term actions should be in order to improve the business outcome.All you need is a few charts with great data visualization – and this is worth 1,000 slides.

Picture1Picture2

The next good tip is –  keep reports easy to understand and don’t forget the actionable steps that need to be taken based on the visualization insights or report insights. Your data visualization should always include recommendations as to what the business user or decision maker need to do.

It’s a 4 step process, ‘DATA – VISUAL INSIGHTS – ACTION – MEASURE’

When it comes to Big Data, it’s important to ask the right questions about what kind of information can empower your business, and your customers.

Identify trends and looking at what people are looking out for during that period. Information can be rendered useless or useful during different periods in your customers’ life.

For example, we found 35 out of 450,000 customers who were at high risk of leaving. That’s a small number, but the loss of those customers would have meant a loss of about $3.75 million dollars.

Picture3

If the company had waited for a completed data warehouse implementation, this insight would have been missed and the company would have been in danger of closure!

Data Scientists are Making Healthcare More Effective

doctor-showing-result-to-patients-18076806

I was recently inspired to write this article below, when I read a long while ago about how patients are most times treated as some sort of ‘average’.”Typically, a patient will receive a treatment based on what worked for most people, as reflected in large clinical studies. For example, for a long time, it was thought that Tamoxifen was roughly 80% effective for breast cancer patients. But now we know that it is 100% effective in 70% to 80% of the patients, and ineffective in the rest”.

Today, there is a huge opportunity for Data Scientists to put their Predictive Analytics / Statistical skills to work, as today, we have access to a new kind of data such as ‘DNA sequencing’ to tell whether it’s likely that a drug will be effective or ineffective in any given patient, and we can tell in advance whether to treat with the drug or to try something else.

Today, we can use predictive analytic techniques such as the Logistic Regression or Decision Trees to divide patients into groups and then determine the difference between those groups. With the Logistic Regression Modelling, we can tell who is likely to be cured with a particular treatment and also, the probability of being cured with that particular treatment. Decision Trees also offer good visualisation showing the breakdown on the different segments of patients who are likely to cured by a particular treatment or not.

Many focus on whether a treatment is effective or not. The fundamental question is, “for which patients is this treatment effective”? It’s all about asking the right questions….The question should always include the patient not just the treatment! A treatment that is only effective on 25% of patients might be very valuable if we can tell who that 25%  is.

One of Data Science’s many promises, is that, if we can collect enough data effectively, we will be able to predict more accurately which treatments will be effective for which patients and which treatments won’t.

If you are interested in “What treatments will work and on which type of patients?” or “Whether you are spending money effectively on your treatment?” Why not get help from Data Scientists, like myself, who can apply Predictive Analytics / Statistical Techniques to historical patient and treatment data and help you answer such questions.