Contextual Targeting – More than a Cookie Substitute: Q&A with Silverbullet's Marco Godina

With the sun rapidly setting on third-party cookies, a persistent global pandemic, and social and political unrest around the world, there has never been more pressure on brands to ensure that their advertising is not just accurate, but relevant and sensitive. However, the ad industry has proved that it is willing and able to adapt, with firms exploring new solutions to power an ecosystem that serves consumers and companies alike. One such solution that has seen a resurgence amidst tightening privacy regulation is contextual, which offers a more privacy-centric targeting solution by using a website's contents to determine which ads are served to each end user. Marco Godina (pictured below), SVP of product at Silverbullet talks exclusively to ExchangeWire about the benefits of contextual targeting, and explains why the solution is more than just a placeholder for the near-extinct third-party cookie.


In the wake of the COVID crisis and new privacy regulation, brand safety has become more critical than ever. How do you think industry players are faring when it comes to brand safety, and what do you think they could be doing better?


COVID-19 has really shone a spotlight on how limited some of the current approaches to brand safety are. For starters, much of the industry is still running on simple blocklists, or using outdated technology that only spots and flags individual keywords without assessing their context. Unfortunately, this only serves to misinterpret, or misrepresent what an article is really about – such as instances where articles about Meghan and Harry are blocked because ‘Sussex’ contains the word ‘sex’. Now, in a climate where almost every article is about the global pandemic or political and social unrest, these approaches result in a huge chunk of quality content being unnecessarily dropped, essentially leaving brands with limited places to engage their customers, and publishers with limited ways to monetise their inventory.

Another big problem, which often gets overlooked, is that advertisers are forced to approach brand safety differently based on which channel they’re executing on. The safety and suitability signals available within walled gardens are very different from the signals available on the open web for display - which in turn are very different from the signals that can be used on programmatic video and programmatic audio. As the space evolves, I personally think we’re going to start seeing advertisers demand parity with the way brand safety is managed and executed across all of their digital channels.

On the bright side, we are starting to see newer contextual solutions adopt things like natural language processing, and other advanced techniques that go beyond keywords in order to understand the semantics, context, and sentiment of an article in the way that a human would. We’re also starting to see new tools that are applying innovative approaches to improve understanding of video and audio as well. This is the direction the industry needs to move in – using smarter and more nuanced solutions to help brands find suitable content to advertise alongside (regardless of inventory type), whilst helping publishers monetise meaningful and reputable content.


With third-party cookies nearing complete extinction, contextual targeting is experiencing a bit of a resurgence. How could contextual serve brands under tighter privacy regulations?


Contextual advertising has become sexy now, right?! I’ve been in ad tech for over 15 years, so there’s an odd sense of déjà vu seeing everyone circling back and innovating on some of these traditional digital advertising techniques.

Before I joined Silverbullet, I worked at Bluekai (and then Oracle), so I had a front row seat to see what happens to third-party data when privacy regulations become a reality. In short, the third-party data world is dead. It’s game over. Advertisers have to find a new way to communicate with their customers, and context is a very safe and privacy compliant way to do it. Why? Well, it's executed pre-bid and it doesn’t rely on cookies, making it amenable with all the current privacy legislation and browser requirements.

Currently, a lot of brands and agencies still think about context as a solution that they only use to help manage their brand safety on programmatic buys, but I believe we’re going to see many of them adopt context to fill the reach gap left by third-party cookies. I personally think that context is going to deliver that reach in a more transparent way, without the uncertainty about quality that always plagued third-party data.

The reality is that third-party cookies and their associated data weren’t great. These data assets helped advertisers gain reach in a semi-intelligent way, but their weakness was always accuracy – too much obfuscation produced poor data quality, which couldn’t be verified in a meaningful way.

Let’s take a look at an example: a third-party cookie tells me a consumer is ‘Male, with an income of £100k a year and is in-market for an SUV’. As an advertiser how can I verify that information is true or accurate? But on the contextual side, this is much easier to assess: if a user is currently reading an article about the ‘Best SUVs in 2020’, it’s safe to say that user is highly likely to be in-market for a new SUV.

As providers continue to innovate the tech behind contextual targeting solutions, the approaches and methods of engaging with customers will only continue to become more sophisticated and meaningful.


What are the greatest challenges for brands looking for a brand-safe environment, and how can contextual help them to overcome these?

Marco Godina, SVP of product, Silverbullet

I’m sure most will agree that we’re in a weird place right now. 2020 has seen a pandemic literally shut the entire world down. Racial inequality is meeting a due reckoning, with protests spreading from the US to most countries across the globe. If these weren’t enough, we’re also living in a highly polarised political climate where it’s difficult for people to tell when information is objectively true or false.

How does a brand approach brand safety in the current climate? That’s the biggest challenge right now – when most of the content on the web is tangentially related to these world-altering issues, what does a brand do?

The short answer is that it acts with nuance and shifts its mindset. Each brand needs to assess their approach based on their own brand profile and appetite for risk, but sticking to their old brand safety tactics alone will no longer be effective. If your brand decides to avoid all mention of ‘Coronavirus’ in digital, then you’re going to quickly encounter reach issues. This will only lead to you falling out of touch with your consumers, paving the way for a competitor to take your place.

The best solution is to use good contextual targeting to start shifting away from the legacy black and white concept of brand safety, and move towards advanced brand suitability tools that help factor in more nuance around when it is and isn’t appropriate to advertise.

Good contextual solutions can help brands understand things like the sentiment of a topic. Sticking with the example of the pandemic, not all content that mention ‘Coronavirus’ is bad – an article about Tom Moore raising £30m  is very different to one about the rate of COVID-related deaths soaring in Arizona. Other solutions can help explicitly avoid content that’s more damaging and difficult to detect, such as fake news and racism.


With the deprecation of third-party cookies, first-party data is becoming hugely important to brands and publishers alike – how can brands master first-party data?


At Silverbullet, we believe that the death of third-party cookies is an opportunity for the industry to change; to place the consumers at the heart of engagements, and to innovate in a privacy-compliant way.

To succeed in this new era of marketing without third-party cookies, it’s imperative for marketers and their technology partners to be willing to evolve how they source, manage and implement first-party data – this truly is key for the future.

If brands want to continue creating data-driven customer experiences in a privacy-conscious world, they should begin to explore alternative methods of creating a frictionless flow of data to fuel and inform their business. A few alternative methods brands should start to explore when it comes to first-party data include:

  • Innovation around generating consent-driven first-party data pools
  • Deploying technology that can effectively manage data and distribute it across all channels that matter
  • Maximising use-cases driven by first-party data
  • Leveraging solutions that can translate first-party data into contextual targets and other future-proof solutions
  • Working with premium publishers to increase relevancy and, hopefully, incentivise consumers to willingly share their data.


What is Silverbullet doing to help brands and agencies in a post-cookie, more highly-regulated landscape?


Silverbullet was founded four years ago to help advertisers figure out the best way to leverage their first-party data. In 2019 we realised our mission needed to evolve to include additional products and services to meet our client needs, so we founded the Silverbullet Data Services Group. To that end we’ve just launched our newest product, a contextual intelligence solution called 4D, a tool that addresses a lot of the gripes we’ve been hearing from our enterprise customers.

As the data space evolved, particularly with the introduction of GDPR and CCPA, Silverbullet noticed that our enterprise clients were spending massive amounts of money on SaaS contracts to manage first-party data that they subsequently had trouble scaling.

We believe that advertisers can pair context with first-party data to help get the best of both worlds. By ingesting first-party data from the leading DMPs, CDPs, and Ad Servers, our 4D solution is able to give advertisers contextual insights that can immediately be used to inform programmatic targeting, all without touching a third-party cookie.

We’ve also taken a really innovative approach to video in 4D by utilising natural language processing and computer vision to understand the true context of a video and its story. By analysing the unique images and entities contained within a video, our solution lets advertisers (for the first time) use accurate signals to inform their targeting and brand safety strategies in programmatic video.

Lastly, 4D provides a contextual marketplace to give brands and agencies access to some ground-breaking contextual solutions. We’ve partnered with some innovative providers to bring approaches from academia and broader ad tech into the contextual space. A great example of this is our launch partner Factmata. Factmata can analyse and score content using nine signals, including political bias and hate speech, to provide a deep understanding of the quality, safety, and credibility of any piece of content on the web. Clients can use this to inform pre-bid targeting alongside all the other keyword based signals that 4D provides.