Future proofing in the age of machine learning – Roxanne Brownlee, Google – Nottingham Digital Summit

Posted on 03/07/2019 by Team Hallam

The fourth talk on the Albert Hall Stage for the Nottingham Digital Summit is Roxanne Brownlee of Google, who spoke about how businesses can stay relevant through new technology.

Roxanne Brownlee, Nottingham Digital Summit

Why is today different?

Mobile traffic was just 1.5% of total web traffic in 2010, yet that was the year when Google first announced the importance of being mobile-first.

9 years later, mobile traffic has now shot up to almost 60%.

Agility is the difference between surviving and thriving. To thrive, you need to acknowledge what is coming in the future.

Companies who offer value to the customer, and offer a superior user experience before the expectation existed and it became competitive; found it easier to thrive later on.

What it means to thrive online

Technology and the world around it are changing more rapidly than ever. Businesses that can adapt to leverage today’s technologies will have an edge tomorrow.

Technology is moving fast and so are consumers, so brands need to catch up. Data driven marketing has become the key to delivering quicker, more relevant messages at scale.

Technology can empower brands to become more helpful, by letting them be there for their users, with useful answers provided timely and seamlessly.

The ability to predict customers’ needs using data insights and technology will define a brand’s ability to grow. In fact, we’re seeing it already:


Business’ use machine learning to help provide customers with relevant, real-time messages at a large scale, in a much quicker space of time.

Done right, using personalisation to deliver messages tuned to what customers want – or are about to want – is proven to drive engagement and loyalty. If customers benefit, companies benefit.

But like many businesses have found – putting all this into practice isn’t easy.

  • People spend 9 hours a day online
  • People check their phones 150 times a day
  • 70 million signals are collected in the blink of an eye
  • 90% of users switch between screens to complete tasks
  • 62% of all marketers are feeling overwhelmed by the size and volume of data

If we know people are checking their phones 150 times a day, and spending 9 hours online, then we can gain a rich insight into who they are, what they care about, and what they want. This is the kind of data we can bring into our audience strategies to drive relevant and performance in our campaigns.

When we build strategies based on keywords alone, we ignore this data we have. We judge them exclusively on the tiniest snapshot of information, rather than looking at the bigger picture.

Every day, we produce 2,500,000,000,000,000,000 bytes of data.

This translates to one trillion essays of data every single day.

We have all the data we need to create incredibly relevant, targeted advertising. This means we can make really smart and impactful business decisions.

All these fragmented moments of interest and intent generate too much data for humans to process. It’s therefore necessary that advertisers understand these fragmented journeys, and know how to engage with users throughout these moments of intent.

Much of our day-to-day technology is powered by artificial intelligence (AI). For example: we train a software system with lots of examples so it can pick up on patterns. Rather than telling the computer that all spam emails contain the phrase “new weight loss trick!”, you train it on millions of examples of spam, making small corrections until it can pick out the pattern on its own. This ability to learn patterns is called machine learning, and can make your life easier in so many ways.

Some examples of Google’s products that use machine learning in the digital marketing industry, and non-ad related products include:

  • Universal App Campaigns
  • Smart Shopping Campaigns
  • Affinity Audiences
  • Responsive Display Ads
  • Google Maps
  • Gmail
  • YouTube
  • Google Translate

More than half of the world will have grown up in a digital first mindset. While in 2019, millennials make up 24% of the global population, by 2025, 55% of Brits will be millennials or younger.

We are now in the age of machine learning

Advertisers who are successful at data-driven marketing are seeing revenue increases of as much as 20%, and cost savings of up to 30% (Boston Consultancy Group, Digital Maturity Study 2019).

Yet these companies who fully capture its potential to see their revenues grow and costs reduce, will reap the benefits of a huge opportunity.

The same research found that both technological and organisational factors contribute to a brand’s success in its ability to dynamically execute and optimise toward single-customer business outcomes across channels.

We’ve found that the best in class brands excel specifically in three areas:

  • Data-driven audience insights: Building a holistic customer view by linking online and offline data to reach them along the funnel, based on demand spaces
  • Automated activation: Programmatic media buying optimised with cross-channel signals and personalising creative in a co-ordinated & sequenced across channels
  • Measurement and attribution: Measurement including fractional attribution and frequent testing, optimisation of KPI’s in-flight activity

But why should all this matter to businesses? We’re in a unique place when businesses can transform their marketing strategies to reach the more connected consumers, improve business results and more importantly become category leaders in this space.

What you can do about it

According to eMarketer, video and AI are top priorities for marketers in 2019.

Today, machine learning helps drive the best performance for your ads through automation. The algorithm is constantly learning based on real-time feedback, and adjusting to improve performance.

Machine learning is present across Google solutions, together helping advertisers achieve the best performance for their campaigns by showing the right message to the right customers in their moment of need. Machine learning does the work to answer to your marketing needs:

How do I find my ideal audience?

In the world of targeting, we utilise machine learning to find people who aren’t your customers, but should be. Google helps advertisers expand their reach to those people who are likely to convert. We are able to effectively capture intent through our audience solutions.

How can I measure performance?

In the world of attribution, we utilise machine learning to understand the true drivers of conversion and assign the right value. This means that we are able to measure intent better through non-last click, data-driven attribution.

Finally, we can to capitalise on this intent by tailoring our bids and creative ad assets to best respond to user intent. They help answer:

How much should I bid per auction?

With machine learning, we can factor in countless signals to determine the optimal bid for each moment, so you can acquire more customers and eliminate wasteful spending.

What message do I show my audience?

For creatives, Google helps advertisers use the right message for every moment. With machine learning, we can use our understanding of consumer intent and data to choose the right creative to show for every moment.

Focusing on PPC, bidding was pretty simple a few years ago. You chose keywords and specified the bid based on the likelihood of that keyword converting on a last-click, desktop/same session basis.

Then mobile devices came along, creating more signals that marketers needed to account for, such as device type, time and location.

The key to masterful bidding is to adjust your bids based on each user’s unique combination of signals. Asserting all of these manually for every auction is impossible – even for the highest skilled account managers.

Google Smart Bidding is aimed at setting real-time bidding, for each and every auction. As it considers billions of consumer data points everyday – from colour and tone preference on mobile screens, to purchase history, device and location – it’s possible to use the data to deliver millions of ads customised for your customers, and set the right bid for each of those ads in real-time.

Fully automated Smart Bidding solutions drive the best results. While most of us are familiar and have seen the success with eCPC, fully automated smart bidding strategies like maximise conversions, Target CPA and Target ROAS have the largest impact with advertisers.

Through internal product analysis we found that our fully automated bidding strategies have superior performance as compared to our semi-automated solution (eCPC):

  • Search: 7% avg uplift for eCPC vs >20% avg for tCPA or Max Conversions and 35% avg for tROAS.

However, it’s important to note that eCPC is not a fully-automated solution!

Getting smart with data gives you space to be creative

By getting some time back, you can be more creative. You can spend time understanding the changes in your customer behaviour. For example, one big shift we’ve seen is the decline in time the average person in the UK spends engaging with traditional marketing. For example, over the last five years, we are watching 22% less TV, averaging at 2.5 hours a day. Instead, we’re engaging with online media for over 7 hours – specifically on smart phones.

We now have the space to plan for the future today.

The reason why we talk about digital video being cutting edge is because by 2021, we expect 82% of the global web traffic to be made up of digital video.

If we’re not preparing for this in advance and figuring out how can we make digital video work for us, to grow our brands, in short, we’re going to be failing to respond to 82% of the internet. In a few years, the businesses that ignore digital video will be like the businesses that are ignoring mobile today. The video trend presents a key opportunity for brand building in a future proof way.

  • 1.5 billion people are logging in and coming to YouTube every single month. The average person spends over an hour every day, watching YouTube on mobile.
  • 80% of shoppers who watched a video related to a purchase they planned to make, said they watched it at the beginning of the process.
  • 64% of shoppers say that YouTube influences their purchase decisions. It’s not just users who are active on the platform – others find that it shapes the way they think about brands, and their likeliness to act.
  • B2B customers who see a brand’s online video are more likely to purchase the brand in the next six months, pay a premium for the brand, and recommend the brand.

We’ve optimised our ad formats to meet the needs of today’s viewers and marketers.

Future proofing in the age of machine learning - Roxanne Brownlee, Google - Nottingham Digital Summit

Enjoy this article?

Subscribe for weekly insights

Future proofing in the age of machine learning – Roxanne Brownlee, Google – Nottingham Digital Summit

Let’s thrive together

Get in touch to take the first step.

Contact us