60 Seconds Wi...Andy Brown

Following a career that started at Anglia Television back in 1986, Andy Brown is now Chairman of media analysis specialist Kantar Media. He works alongside Jean-Michel Portier (formerly CEO of TNS Media), supporting him in a number of areas including digital development and global client accounts, as well as the general management of associates such as IBOPE Media, TAM India and RSMB. He sits on the Kantar Digital Board and WPP Digital Advisory group.

Q: You are speaking in the session Big Data - Broadcastings New Oil or Digital Exhaust' at IBC. Briefly, which one of the two is it?

A: I am not sure it is a simple dichotomy. I do believe big data may have several advantages that it can bring to the broadcast industry when it comes to providing insight and commercial opportunities. Some of the key benefits are scale (providing granularity); cost efficiency (often data is a by-product of another activity and therefore potentially lower investment); closer to realtime and in some cases collected outside the auspices of market research- which can mean that the user has the potential to respond more expeditiously and possibly respond directly to an individual (as there may be no requirement for anonymity). At the same time there will be a huge amount of data collected that may add very little, or have no use, and in those cases often a sample of the data is as valuable as an audit!

Q: Is it a natural and easy fit for broadcast? Or will some of the modelling techniques developed for other industries have to be adapted?

A: In all industries there is a degree of inertia and legacy of existing models. I think certain parts of the broadcast ecosystem can lift and replicate other industries use of big data (e.g. social media being integrated in TV content), while other elements such as the trading currencies for advertising airtime will take slightly longer to incorporate as the money at risk is significant and the industry sources such as BARB need to attain consensus from multiple stakeholders.

Q: Broadcasters have to fund analytics from limited resources What is the pitch for persuading them to look at analytics as opposed, say, to investing in social media integration?

A: I do not really see a difference between the two. Social media analysis is just one form of big data that will be analysed alongside other solutions, for example viewing information from the return path of set-top boxes. In some cases there may be an element of substitution of cost, where lower cost big data may replace more expensive and rigorous market research. In other words, can I get a quick and reasonably accurate read in realtime as opposed to waiting for the greater accuracy? This might be easier to achieve in certain areas such as content creation and scheduling, but less likely in terms of audience measurement for trading advertising.

Q: What sort of insights have you been able to provide broadcasters with at Kantar Media?

A: If we take Return Path Data from set-top boxes for example: Historically we have been able to provide accurate audience data using peoplemeter technology such as that deployed on BARB. The evolution of RPD data allows potentially more granular analysis of TV viewing activity that could not be measured through the peoplemeter panels. There are a number of markets where integration of peoplemeter and RPD data are being tested.

In the US we have also integrated data from storecards with return path data to provide addressable advertising solutions such that relevant ads are shown to consumers interested in buying the product as opposed to using a demographic surrogate. It shows a huge uplift versus traditional demographic targets and begins to address the old conundrum of which half of my advertising is wasted?

Q: What are the limitations, if any, in being able to understand or predict viewer behaviour?

A: Personally I believe the use of RPD data in TV audience measurement and understanding viewers is inevitable. It has already provided predictive insights for our clients as to where cable or satellite customers may be likely to cancel their subscription. However for more general audience prediction, having been in the industry over 20 years, one has seen various modelling techniques and algorithms tested and evaluated, but nothing has yet proved reliable. Still, factors such as dynamic/competitive scheduling and even the weather have a big determination on audience sizes.

Q: There is some reported unease about the ethical considerations of utilising Big Data. What are the safeguards and where are the moral limits?

A: For a market research company such as Kantar Media, we have two important limiting factors:

a) We are bound by our professional code, the MRS code of conduct, which is very clear about how research data can and should be used.

b) Privacy laws- We have been very active in this area by actually appointing a full time privacy officer at Kantar, in order to see that we are complying with the law.

In summary the trust of the consumer is paramount. For clients, the story may be more about the trust. That is to say something maybe legal, but if the consumer thinks it creepy, what will be the long terms impact on the relationship with that consumer?

Q: What do we need to make more progress in the area? New mathematical modelling? Or more data fields to analyse?

A: I think it is an element of both. There will be new techniques applied to data, however in the digital age with greater connectivity across multiple devices in the home and services such as Twitter and Facebook being used to log in, then a growing number of fields will become available and utilised.

Q: You are a big Arsenal fan. How could Big Data currently help Ars ne Wenger?

A:Yes I have this misfortune! I am slightly concerned that Ars ne would use this to find more obscure 16 year olds playing in the reser

LINK:	http://www.ibc.org/page.cfm/action=library/libID=2/libEntryID=206/list...
	See more stories from ibc

Most recent headlines