Product Talk # 16: Big Data and Product Development

On Thursday, 6th of June a ’40 strong body’ of die-hard product management professionals descended upon the Erskine Street Gallery, Sydney for the 16th in a series of Product Talks.

An esteemed lineup of panel speakers included Google’s Kate Ryan, Sean Richards of Pitney Bowes Software, Christian Bartens from Datalicious and our own Nick Coster.

Here is a summary of the Q&A.

1.     What does Big Data mean to you?

Kate’s travel industry clients make better product decisions through data mining. A New Zealand campervan example was provided for improving sales in an underperforming market.

Christian agreed that data analytics is best leveraged to quantify an opportunity and help prioritise the focus of your business.

Sean stated that ‘Big Data’ is not new. It’s certainly the latest buzz word and there is presently lots of hype but the concept has been around for a long time.

Nick boiled it down to asking the right questions and then obtaining the data to validate the answers received. He felt there to be a general underinvestment in using the data to answer questions and a lot of upside for better use of big data.


2.     What are some of the guidelines to using Big Data? Are there any pitfalls of using Big Data that you would like to share?

Nick encourages us to embrace wrong. If Edison hadn’t failed 999 times we would not have the light bulb.

Christian approaches the task as a scientific theory which he then attempts to refute using data. He recommends using your website as a test bed. People are afraid of being wrong but this should not be the case.

Kate says that it is best to take a smaller risk by first testing with a small group of strangers who will be very blunt.  Try website www.usertesting.com for some good usability testing templates .

Sean builds a usage profile of customers of products. His company helps their customers to grid their markets eg.  ‘Lost cause’, ‘Persuable’, ‘Sleeping dog’.

Christian conducted a poll and found that among the larger organisations represented in the room it was only Optus who had structured their CRM  & Product Development teams both reporting up to the same Department Head. A key issue blocking effective use of enterprise data according to Christian is the fact that many organizations operate silos that separate the data generated by their Customers and the Product Development team.

Sean agrees that CRM data is valuable and his company recognised this by having CRM owned by the Marketing Team.

 

3.     How can smaller organisations tap into Big Data?

Nicolas (audience member) representing a small business asked the Panel for advice on identifying patterns in the data set. He acknowledges that Big Data is cheap & easy to get. But what do you do with it? What is the best way to predict trends, recognise patterns and metrics/algorisms that alert people.

Nick Coster predicts that machine based learning will likely evolve as a cost-effective way to identify patterns of behavior that may be leveraged by small business in the future. In the meantime he feels that Product Managers can do better in terms of identifying predicators of products and marketing campaigns.

Christian suggests that firms who are serious about making sense of the vast quantities of enterprise data available should consider hiring Science graduates.  Mathematics as a discipline, he felt, is too tunneled. However a Scientist with a PhD in Physics, Chemistry or Biology has demonstrated that they can think outside the box. They are used to doing their own fieldwork, data gathering and analysis to support or refute theorems.  This skill set is crucial in dealing with dirty data.

Nick spoke about certain decision making that does not require 100% certainty and gave the example of  fraud detection systems using patterns of confidence to find the fraudsters eg capturing IP addresses and analysing the relationships with fraud networks.  We need to accept that it is rarely a perfect data set but we are essentially looking for patterns.

Christian stressed the importance of investing to make the data clean i.e. getting the process of planning  maintaining right so that we make sure the data is clean.

 

4.     How useful is Big Data in Product Development?

Sean sees industry interest and a drive for change by consolidating a single customer view i.e. organisations have lots of legacy systems generating disparate data sets.  He uses organisational  relationships to build a picture and believes companies should review their processes first before they look at their technology. Big Data as a Service has value if we get the right people to talk to us about pooling departmental data.

Natalia Salzberg from Macquarie Uni believes the principles of quantitative research are still the same i.e. You need to have a question (i.e. what are you trying to find out?) and perform research against that question to support your hypothesis.

 

5.     Can Big Data be used for social good?

(i.e. Can the data that people share online directly and indirectly be used to benefit them as the end-user rather than purely for advertising and marketing purposes?)

Kate  referred to a research project by LAPD/Anthropology Dept of University of Southern California where they reviewed 80 years of stats to predict crime hotspots and enabled the police to crackdown on crime.

She also mentioned publically available data making it easier to digest eg. climate change, number of people dying in Syria.  Learning institutions and governments are now providing that data. There is increasing public accessibility eg wikileaks. These are exciting times.

Christian noted that data is being used to predict earthquakes and establish predictive trends eg. outbreaks of influenza .

Stephen Long (audience) asked if Big Data could also be used for evil.

Nick Coster raised the issue where our online experience leaves a trail of activities that could be used badly eg to track our behaviour, put up a fraudulent website, identity theft etc.

Kate mentioned that the Arabspring highlighted the dangers around sharing of info through social media. She said that there is a naivity in the community that only good people are using social media

Christian had read about a recent UN report claiming that big data is culling privacy & free speech. Governments are now arresting people based on their behavior and surveillance data on that person’s behaviour.

Sean mentioned that a data breach has to be published. However Kate said there exists different laws globally and that in general social media can collect data on people without them knowing it.

Nick drew on his experience at EBay receiving enormous quantities of data in and that modern day consumers increasingly accept the fact that their privacy has been compromised.

Adrienne Tan (audience) brought the discussion back to the use of Big Data  and where Product Managers, particularly in smaller companies, can access it.

Kate suggests that smaller organisations can look at Google Trends to review popularity of certain searches. Also the Experian Hitwise service gives an idea of how online data captured may be reported. She perceives greater agility in small organisations to get data. The nimble players may even change their business model according to where the customers are. Annual reports are another useful source of research data for Product Managers.

Sean commented that the Government publishes lots of useful information on various markets.

Christian believes in investing time to collect data on the competitor offerings and ongoing monitoring of this data set.  One website that Christian highly recommends is www.80legs.com for  monitoring competitors eg price changes and product features

He also encourages Product Managers to use data to quantify the opportunity and then prioritise what to spend time on. eg ABS census data can be used to help quantify the size of the market.

What great advice from Christian!