How to Think About Data

When you’ve been working around publishing and insights as long as I have, every pitch comes with the promise of data for measurement and targeting... First-party data. Proprietary data. Third-party data. Syndicated data. BIG data. We are drowning in a sea of data but most people can’t really explain what it all means. It reminds me of an article around the eulogy of the word 'culture' that laments the loss of the word’s meaning and influence.

Is 'data' destined for a similar fate? Maybe not.

Data at its core is just human intention and behaviour collected in a database, and if carefully and thoughtfully extracted can provide rich insight into how and why people choose to spend their time, and inform how to deliver on or enhance those experiences. 

At Vox Media we take a privacy forward approach, we also acknowledge that computers and algorithms fall short because humans are not a flat data point or even a collection of data points. We created a framework for our data suite which allows us to draw insights from the audiences’ contextual experience, which creates a more dynamic understanding of what they want and need from us. 

Start with the consumer: Knowing a person is about far more than objective facts 

We can understand what a consumer is doing through behaviour (which is important), but why they come to us is even more critical. 

In a recent Ted Talk by Timandra Harkness, she challenged whether computers and algorithms could ever “get a person". She said that during the Age of Reason, people explored the subjective mind in order to create true understanding and meaning. My favourite quote from her talk is: “When we say we know a person, it’s not just about facts. To be familiar with a person, a machine does not have subjective knowledge of a person.”

This truth must be applied to how we make sense of data in our day-to-day work: meaning and understanding aren’t based solely on the objective facts of a person but rather on the subjectivity of what the creator intended and why the consumer is there in the first place. 

Currently, digital advertising looks into a microscope to define users instead of leveraging a telescope to understand the full context. At Vox Media, to get a holistic view of the relationship our networks have with our audience, we use a combination of editorial insights, first party data, and custom surveys and intelligence tools to understand their interests, motivations, and behaviours. 

We strive to know a consumer, versus targeting them based on just what they look like in order to better serve their needs. Saying we know a consumer from just behavioural data without context does not work!

Understand the context: Zoom out to include sentiment, intent, motivation, and passion

If you were to have a conversation at a church, a coffee shop or a bar, that context would inform everything — the physical environment matters. In the same way, the digital environment matters. This means two things: not all data is created equal and how you mine data makes all the difference. 

Contextual targeting runs on basic data points we gather about users. It starts with how we might define a user demographically, such as a group of keywords found in the content they consume or by something a user puts in their shopping cart. This premise is basic and mechanical. It’s time to redefine the concept of contextual targeting.

Vox Media is home to seven authoritative networks with loyal, passionate audiences who care deeply about the world. Our audiences come to us with intent, not on a passive scroll. They’re looking for a restaurant recommendation, to understand a complex topic, to discuss their favourite team, buy the best home or tech product, and much more. CRO Ryan Pauley recently discussed this broader definition of context, explaining that to earn audience attention, we must “do a job for audiences. Don’t just look for a transactional relationship based on personal information. Consider the whole context (and the whole person!), and meet audiences in the right mindset." 

When quality content is matched with people’s passions and interests, it creates a much richer picture of who the audience is, why they are there and what motivates them. It also allows us to create personalised experiences and products with intelligent optimisation tools such as Creative Intelligence, Crescendo, and the most responsive algorithm — the human mind. 

When we do this right, we no longer base contextual targeting on just the individual, an individual action, or an individual piece of content, but on what our platforms deliver for the audience holistically. It changes the retroactive perspective of targeting and makes context targeting future-facing. It is preemptive targeting as opposed to retro-active targeting. 

As a data 'anthropologist', this construct allows us to challenge ourselves over the next few months to get even smarter about context and tackle tough questions for all of us. 

Questions such as:

● Does the sentiment of a piece of content matter more than endemic context/placement — and how do we action on that? 

● How should we dissect and organize the publisher experience, to action on first-party data more strategically?

● How do we build a contextual knowledge graph that isn’t just based on keyword grouping, but something larger? 

● Do we see a difference in ad performance based on where an ad is placed across our site?

Measure the connection: Translating authentic connections into advertising results

The third C is earned by truly understanding why a consumer chooses to spend time with us and engaging their passions or needs in the right context. This is connection - an ongoing feedback loop and cycle of real-time personalisation that spans an entire ecosystem of digital content, podcasts, live experiences, and more. The result, for both audiences and marketers, is impact.

Now, how do you measure it? How rich was the connection that you formed with the audience? 

Media consumption is changing all the time and we must have the right solutions and partnerships to help marketers evaluate that connection explicitly. At Vox Media, as we continue to develop our lines of business, we are investing in building and utilising the right data tools to maximise our partners’ ROI. After all, with two billion ad impressions served on our Concert network of 70+ premium publishers this year, there’s a lot to learn, measure and optimise. We also stimulate the auditory senses with podcasts and activate real life experiences through conferences and events.

One recent example took place at Code Conference, which brings together tech luminaries, CEOs, and marketing leaders. One tech partner saw a tremendous opportunity to reach this influential audience as one of our premier sponsors that was both live and digital — and we had to parse out the tangible impact of this cross-medium experience. By partnering with Moat, we were able to show that Code drove massive engagement with video completion rates that were 80% higher above benchmarks we normally see and interaction rates with their ads blew past industry benchmarks by 7X. To top it all off, we also delivered incremental reach in the tens of millions across LinkedIn, Facebook, Apple News, as well as our podcast extensions of the event. 

This authentic and trusted connection translates into for more effective advertising, and measuring it is proof of that.

Challenge ourselves: Letting it guide our decisions

In doing this for such a long time, I can always rely on research and data to keep me humble. There’s always something to learn, something to reframe and something to challenge us to keep getting that great return or innovate to find alternatives in learning moments. 

For example, Nielsen has told us that our branded content rocks. On average our branded content produces +28% lift in terms of uniqueness and average time spent on Vox Creative is 89% higher than their industry norms. But I can tell you that our success is a product of introspection and iteration. 

We are answering tough questions that advertisers ask us, such as: Is branded content more effective in driving down-funnel CTR? Well, it depends. For a finance client, we saw a ~25% increase lift in down funnel conversions when you pair display with branded content. To push this even further, we created and tested multiple pieces of branded content. We saw that when display was paired with the best performing piece of branded content, it generally performed better than display that ran with multiple pieces of branded content, showing - again - that the quality of branded content matters. 

Small things like this matter because we aren’t chasing impressions and frequency, but an advertising experience that is additive and incremental in creating the best consumption experience possible.

So, what now? Start a conversation

None of this is easy, but it’s exciting. Vox Media started as a network of authority, recommendation and activation. We have the voices that move you to do something more: know more, buy more, or even do more. We believe we have the ability and scale to mobilise intent in the most premium places with the most premium ad formats. We want to continue to find new ways advertisers can capitalise on that intent through the unique context we bring to that consumer.

Targeting needs to move away from 'who you think I am' to 'why I am the way I am' within the context of some of the best experiences founded on digital - which we believe we have built and continue to strengthen across Vox Media’s portfolio of networks.

In closing, my last words on data are simple: if you don’t know what you are looking for, you really don’t know what you are looking at. Data isn’t evidence that closes on a conversation with a consumer, but it really is the start and opening of a true and meaningful connection.

Edwin Wong is Vox Media’s Senior Vice President, Insights & Innovation