DIGITAL MEDIA DIGEST: OCT ‘17
A monthly look at the world of digital from NORTH’s point of view
Amazon Alexa Adds Voice Recognition
Caroline Desmond, Director of Media
Last Wednesday, Amazon announced that it’s virtual digital assistant (VDA) called Alexa will be able to recognize different voices. This comes on the heels of new Google Home products and integrations also announced last week. This latest announcement from Alexa is a big deal considering Google Home has enabled voice recognition since last April. The key benefit both Google and Amazon address through this feature is the ability to deliver personalized responses depending on who is asking the question without requiring a device for each user.
Voice recognition will work on Echo, Echo Dot, or Echo Show, but voice recognition is not yet available on the Fire TV remote or the Amazon Tap. Setup requires users to repeat a series of phrases so that Alexa learns differences in voice intonation to be able to differentiate between each member of the household. In doing so, Alexa will be able to summon the unique calendar, playlists, contacts, etc. for each profiled user upon request.
Currently, Google Home and Alexa lead the pack for connected home device VDAs, and each company’s investment in personalized voice search is an indicator of the expected growth for this channel. In fact, shipments of Google Home and Amazon Echo speakers are expected to climb more than threefold to 24.5 million in 2017, according to a report from VoiceLabs (the company that Amazon partnered with to allow third-party developers to build skills for the voice assistant). Furthermore, Tractica, a market intelligence firm that focuses on human interaction with technology, forecasts that VDAs will reach 1.8b users worldwide by 2021.
With this potential growth, we are advising brands to begin thinking about the impact of VDAs on channel strategies. Some considerations include:
- First impressions are more important than ever as repeat purchases become more automated. On this note, younger audiences hold greater opportunity for brands who can establish themselves early on as the trusted go-to for someone entering a product category for the first time.
- Putting VDAs in control of ordering means greater potential for consumers to view day-to-day products as commodities. Commands are more likely to be generic (Alexa, order detergent) and Alexa ships the best detergent on deal. This makes it crucial for brands to differentiate to create preference.
- Implement structured data on your website. This way, Google can better recognize content on the page and pull from it to answer relevant Google Home voice queries.
- Invest in Amazon Choice status. Brands with strong Amazon presence should invest in Amazon Choice status to increase the likelihood of ranking in Alexa voice search results. For those focused on getting into the Amazon’s Choice program, managing product reviews is the most important factor.
- Test connected home device targeting. Media companies like Pandora and Hulu provide connected home device targeting for paid media buys. This enables brands to stay top on mind in close proximity to voice search purchase opportunities made at home.
New AdWords Feature: Data-Driven Attribution
Devon Brown, Performance Marketing Manager
Search behavior is becoming increasingly more complex. This video from Google explains that a person performs an average of 34 searches across multiple screens for a travel purchase, and 139 searches for an automotive purchase. To gain a better understanding of this purchase journey, Google introduced data-driven attribution. This is a pivotal feature that will change the way multi-touch channels are measured and interpreted. If you’re already confused, let me explain. Attribution modeling is defined as “the rule, or set of rules, that determines how credit for sales and conversions is assigned to touchpoints in conversion paths.” -Google
To demonstrate, if you have three channels in your media campaign, and a user views channel A, then views channel B and goes to the website, then views channel C, goes to the website then converts, the attribution model dictates how each of those channels gets credit for the conversion. In a linear attribution model they would all get equal credit. In a first touch attribution model, channel A would get 100% of the credit. In last touch attribution, channel C would get 100% of the credit.
Up until now, all of these models only measured people who converted. With data-driven attribution, it looks at people who converted and people who did not, and identifies patterns and other behaviors that indicate intent and preference changes. By switching to pattern and behavior predictions, a few key benefits are to be gained:
- Learn which ad groups, keywords, and creative play the biggest role in changing audience preferences
- Understand which components of your campaign impact conversions directly or as “assists”
- Remove uncertainty and guesswork within current attribution modeling
In order to access this feature in AdWords, there a few account minimums that must be met:
- The account must have 15K clicks
- Must have at least 600 conversion actions within the last 30 days
- Once the above requirements are met, the model will start collecting data immediately, and data will be viewable after 30 days
- If data drops below 10K clicks or 400 conversions within 30 days, data will become unavailable
My advice- set up a lot of conversions. Make every possible on-page click, consumer path, or event a conversion to be eligible for the data. You can always optimize toward just one, and no matter what, you’ll gain valuable insight.
As we get our clients ramped up on data-driven attribution, we’ll be sure to update with our thoughts, tips, and insights.
Chatbots. Tell me more.
Izzy Kramer, Media Planner
By next year, Gartner is predicting over 30% of interactions with technology will be conversation with smart machines, including chatbots.
But what is a chatbot? Robots that communicate either by text and in a way that is advertently human. Chatbots fall into a category of artificial intelligence. Think Wall-E but more virtual and less trash compaction.
Chatbot complexity is on a spectrum from simple to advanced, with the more advanced chatbots being able to understand and react to conversational and colloquial phrasing. They are learning machines that grow and advance the more conversations they have. It is these more advanced bots that are intriguing to brands.
Chatbots are extremely useful in communicating basic and fundamental parts of a business. For example, if a customer uses the chat feature on 1–800-CONTACTS to ask how they can send in their prescription, a chatbot is waiting on the other end instead of requiring a person to wait idly in that chat window. It allows companies to increase efficiency and use employees where they are more useful, in tasks that require more human interaction.
Chatbots also benefit consumers who require easily accessible assistance whenever possible. At the end of the day chatbots, reduce the workload for workers and meet consumer demands 24/7.
Despite the listed pros above, chatbots are still a daunting idea. Computers that learn and evolve from human interaction overtime? Sounds like something out of Bradbury fiction. Ummm . . . freaky? And with that said, chatbots are not breaking news. They have been around for years now.
So, what has been interesting as of late is the reactions and opinions consumers have about chatbots now that they are have been sufficiently active and growing in the world.
According to a study done by LivePerson (ironic considering we’re talking chatbots) that was reported on by The Next Web, “38% of people surveyed felt positive about their experiences [with chatbots], while only 11% felt negatively … an overwhelming 67% of those surveyed used a chatbot for customer support in the last year,” and “14% have used one to help with productivity.” To summarize, people don’t hate chatbots, and more importantly, they kind of like them.
Furthermore, chatbots have become so inculcated into website UI that now it is odd (and generally frustrating) to be on a site without one. When consumers are offered immediate and reliable customer service and then do not find the same service on other websites, it reflects poorly on the brand that has not taken the steps to introduced their own chatbot.
Due to their positive ratings, chatbots have also started to branch out from retail customer service. Mostly available in mobile apps, chatbots are helping out when it comes to finding food near you, figuring out your personal finances, and even providing law advice!
All chatbots, with these apps as examples, are housed in messaging platforms either in the app or more recently Facebook Messenger. It is this one-on-one, personalized messaging that makes it feel like you’re talking to an incredibly attentive and helpful friend. So, it is no wonder why chatbot popularity has grown as they have developed over the years. There is no doubt chatbots will stick around and will be given even more room to grow. The question is how will chatbot technology continue to be used by brands or implemented into other technologies?
The future is here, folks!
YouTube Announces Next Level Personalization
Sean Brennecke, Assistant Media Planner
YouTube just brought out the big guns. Google, which owns YouTube, recently announced a series of new tools for YouTube that will help advertisers develop, target and personalize ad creative in an entirely new way.
Let’s dive into the first tool:
Director Mix
Director Mix is YouTube’s latest effort to help advertisers personalize and optimize their video campaigns. The tool enables advertisers to generate hundreds — even thousands — of unique video ads tailored to different audience all from a single creative asset.
Here’s how it works:
- Advertisers uploads a variety of different voiceovers, backgrounds, copy and campaign details (e.g., targeting, creative matrix, ad parameters, etc.)
- Director Mix generates thousands of different versions of the video ad to match your various audience segments
- Advertisers can easily review and approve video creative within the tool
- Advertisers can then upload videos to their YouTube channel and full campaigns to AdWords for launch
Through machine learning, Director Mix can personalize ad messaging to make it more relevant to the content that people are about to watch. Campbell’s Soup has already tapped Director mix to create a series of bumper videos with copy adapted for people watching clips from the show “Orange is the New Black.” For example, viewers would see the tagline “Does your cooking make prison food seem good? We’ve got a soup for that.” Additionally, viewers that watched Beyonce’s “Single Ladies” music video were shown the tagline, “Dinner for one?” noted AdAge. And the initial results look promising. Campbell’s earned a 55% lift in sales and a 24% lift in ad recall with this campaign, reported Google.
All in all, Director Mix could help advertisers improve ad relevancy and reduce the labor and costs associated with developing multiple ad variations. But a couple caveats before you go all in:
- Don’t rely on AI to create thousands of ad variations just because you can. Develop at least one salient version of the ad to ensure the core message is in place. Start small and always have a real person QA and approve each variation
- You get out what you put into it. This may seem like a no-brainer, but be sure to upload quality creative assets
- Use the technology when and where it makes sense. Director Mix will likely work best with performance-based campaigns, or when testing messaging with different audience segments
Director Mix is still in alpha and is currently only available in the US. It plans to roll the tool out globally in Q4 with all languages available.
Custom Affinity Audiences
Google recently announced that it is expanding the capabilities of Custom Affinity Audiences to provide an added level of targeting and precision. Previously, advertisers were only able to target users based on Google search data. Now, advertisers can target people based on the mobile apps they’ve downloaded, as well as the places they’ve visited or searched for using Google Maps data.
In an example from Google, companies such as an outdoor retailer could use this technology to target an ad specifically to skiers by seeking out those who searched for skis, recently visited a ski resort, or downloaded a ski resort’s trail guide app.
Here’s why this is significant: Google reports that by using intent-based audiences on mobile, you can achieve 20% higher ad recall lift and 50% higher brand awareness lift versus campaigns based on demographic audiences.
Despite whether you think it’s creepy or cool, it’s time to explore ways to harness this data to deliver relevant ads more efficiently.
Video Ad Sequencing
Google’s new Video Ad Sequencing feature in AdWords Labs enables advertisers to design an ad experience that unfolds over time. Let me explain. Rather than serving viewers the same video over and over, advertisers can now string together a planned sequence of videos to tell a story. Furthermore, each viewer will experience different paths depending on the ads they have engaged with.
Google mentioned that Ubisoft used this tool to promote its upcoming “Assassin’s Creed” game by serving viewers 6-second bumper ads that featured elements from the main trailer. The campaign was a success as it reached 15 million unique viewers, increased awareness by 25% and search life for “Assassin’s Creed” by 224%.
Introducing the Video Ad Sequencing feature was a smart move by Google and a win-win for brands and consumers. Brands are able to tell their story in a more creative and compelling way, which provides more opportunities for consumers to learn and connect with said brands.
