How to Use Data to Drive Your Media Business Forward: Q&A with Lutz Finger of LinkedIn
Content-centric media companies, including legacy ones, can be successful, profitable businesses, provided they are open to new opportunities and smart about how they go about developing these opportunities. Big data, as often said in past years, holds one of the keys. “The promise of big data is making the wishes and habits of customers visible so they can be acted on,” says LinkedIn’s Lutz Finger, a self-described “Big Data Guy” who combines his background in quantum physics with his passion for analytics, data and new media.
Lutz “likes to make Big Data more accessible” and is the co-author of “Ask, Measure, Learn”, a non-technical book on how to extract significant business value from big data. He will speak at the Digital Innovators’ Summit in Berlin on 23-24 March this year, where he will present on the topic of how to gain value from big data in a content business. He will also provide guidance on how to build a corporate data team in a separate, more intimate break-out session later in the day.
An authority on social media analytics, Lutz currently holds the position of Director Data Science & Data Engineering at LinkedIn and is a co-founder and former CEO of Fisheye Analytics, a media data-mining company that was acquired by the WPP group.
We spoke to Lutz about big data and content-centric media businesses, as seen from his personal perspective (rather than from a LinkedIn perspective) as co-author of “Ask, Measure, Learn” and a big data expert.
There has been a lot of hype around Big Data. What do you think is truly important when it comes to data?
The word hype is used correctly about big data. Big data is not what anybody wants. You want to make big data small or actionable. What big data means for business is that thinking can be delayed. Previously with Business Intelligence (BI) systems, you had to do the thinking up front. There were always the questions of what exactly you wanted to do, what you wanted to achieve, what you wanted to display, what should be measured and so on. People and businesses had difficulty with this but they agreed on a structure, agreed on a schema, agreed on a database and the things that they wanted to measure. The process typically took forever and they got an answer but often realized it was not the answer they were looking for, so they would ask a new question and start over again.
The only thing that big data has changed is that you don’t need to ask the questions up front – you can think afterwards. You collect and save all the data, structured and unstructured. It surely helps to show the same discipline as in traditional BI and know what you want to answer to ensure that you are collecting as much relevant data as possible. But the actual process of posing the question can and often is delayed.
How can media companies benefit from big data?
The promise of big data is making the wishes and habits of customers visible so they can be acted on. This often comes down to the technology of recommendation engines.
The media business consists of the distribution of media, the creation of content and the marketing of content to sell ads and subscriptions. This ecosystem is under fire. Distribution channels have changed and multiplied. Publishing has become cheaper and now everyone can write or curate content. The new distribution channels as well as the new content creators or curators have entered into the advertising business, reducing one of the main revenue streams.
Let’s look at distribution first. Traditionally, a media company looked at where they wanted to distribute their publications and went out and found places to carry their publications. Media companies had a system and data to analyze which shop, which region they should best target. The distribution system has changed from physical distribution to digital but the way we analyze this issue has not. Media companies can still effectively use data to determine which channels are important and how to use them to get the desired reach. Since traditional media companies waited a long time before boarding the digital revolution, the balance of power has shifted towards high-traffic sites controlling the distribution. That said, it is not impossible to build up an own digital distribution channel. Companies like Buzzfeed or Re/code have shown this impressively.
Secondly let’s look at content: It used to be that media companies depended primarily on an editor who had a gut feel for what his audience wanted and he was well compensated for this sense. Companies isolated him from the organizational structures so he could focus on his gut feel to create and select content that the audience would like. Some data, such as circulation numbers, were available but not usually timely or helpful. To determine what content to curate to fit best the audience is still the right business model for media companies but we can better determine what people like based on data. We can now measure many things, such as how far someone scrolls down on an article or how hectic a person moves his fingers over an iPad page, in order to understand what people read, what they skim or where they start to lose interest. This does not replace or make the editor obsolete. The editor still has an important role because the algorithms used for content creation use historical data to be trained. Thus novel inspiration will be hard to generate and the gut feel of the editor helps to create interesting novel stories or ideas in a way that data cannot.
Let’s consider also the cost of content creation. It has dropped strongly, opening up new ways to work with data. For example, I am not a journalist but I contribute content to Forbes for free. The Forbes model shows effectively how to utilize this new free source of content.
Let’s look at the third area: advertising. Distribution as well as content creation and curation has changed the way we can monetize it. Thus the high margin business is gone and media companies need to accept it. However, companies can use data in the new systems to find the monetization opportunities. The data is different but the business questions are still the same.
Some media companies complain that when they create a piece of content and some people read it and like it, Google is capturing the revenue from the action they take. That is true. However, there are opportunities to work with the knowledge in the market. If you, as a media company, know what articles you will be publishing and if you can assess their effects on click rates, you have a kind of insider knowledge that can help you monetize ads more effectively than others who do not have this knowledge.
There are pockets where you can have a successful business. Media companies are too stuck in their old thinking and think they can’t make money.
Yes, the ecosystem has changed, but we as consumers are still looking at content and we are consuming way more then ever before. There are successful media companies in the market. Look at Kara Swisher, the founder of Re/code and AllThingsD. AllThingsD ended because its owner, Dow Jones, could not come to an arrangement with Kara to continue. She left, built up Re/code and became profitable in a short time. Media companies say they can’t make money but Kara did with her business model.
Or look at the New York Times and its stunning adjustment to the new ecosystems. Some believe that the way everyone can create content is destroying the journalism business, but the opposite is true. More than ever before we need to validate information... Is the information true? Can it be validated? Can we put it in context? This used to be the job of the journalist but it seemed to be forgotten in the discussion on monetization. However, this is still a huge business – look at the New York Times and the Economist. They do exactly this and their readers pay them for it.
What questions should a media company be asking before launching into a big data project?
Ask why? The approach is laid out in my book, “Ask, Measure, Learn”. Ask a question, start measuring it and take the learnings out of it. Just digging into data and coming up with a lot of answers to questions people have never asked is not a good idea. To prepare for big data, I recommend that a company do three things:
- Train the management team on data. Let them know what is possible. Teach them to think in data terms. Get them to become data driven. Have them be able to phrase the “ask”.
- Hire people who can work with data
- Store all data and have centralized access to all the data sources. Have the legal team create user terms and conditions that make access to this data possible.
Then when you have a question, you formulate it and then you can structure the data. The most important thing is to determine what you want to get out of it. Having a data mindset and capabilities in your company is important to be able to structure the data in order to make it actionable and to make data driven decisions.
Privacy, especially in Europe, is a big issue. How should companies deal with data privacy laws?
Privacy is a hugely important issue that comes down to a lot of areas of trust and usefulness. You have to be open about how data is used. Many people do not understand yet what is happening with their data. So if they believe that a company is doing something bad with their personal information, it can damage the brand. Transparency is thus the most important word for any company who works with data.
Some media companies have this notion that if they use behavioral data from their readership it will creep out their readers. The reality is probably the opposite. The key is to offer transparency into the user benefits – for example a utility or monetary savings. People are willing to give away their private information for benefits. Those can be sometimes very small, such as the ones a frequent buyer card offers.
Take an example such as iTunes, Spotify, Netflix etc.: Users accept the terms and conditions because they want to use it and because they like the suggestions of other music. Another example is a peer-to-peer lending network that uses a person’s Facebook connections to help establish creditworthiness. Sounds creepy but the approach is similar to what Grameen Bank does, using an individual’s community to establish creditworthiness. While the Grameen Bank founder Muhammad Yunus got a Nobel price, we are scared of the peer-to-peer lending networks. But if those lending networks offer a better interest rate than the bank, most people will likely use them and give up privacy for that.
In some cases, companies are collecting and using data to determine pricing. For example, some auto insurance companies are offering discounts to people willing to put GPS enabled devices into their cars to track their driving habits to show that they are a good driver. Some people are OK with this and others not.
With privacy, we need to be transparent. We also need regulatory authorities to enforce this transparency and help with educational efforts. Privacy concerns can be addressed and they should not stop the possible things that can be done with data.
Come see Lutz and other leading speakers at the Digital Innovators’ Summit in Berlin, Germany on 23-24 March 2015.