Alexa Is Not for Advertising
Why Most Brand Marketers Fail at Building Amazon Alexa Skills — and How to Fix It
Dear Brand Marketers:
We’ve had enough. Please stop.
Please stop trying to build Skills for Amazon Alexa devices. Yes, this is a form of Artificial Intelligence, and you feel compelled to check that box in your marketing strategy. Intelligent assistants are taking off and you don’t want to be left behind.
But what you’re doing is wasting everyone’s time and turning a very powerful platform into a joke. Just because you CAN do something doesn’t mean you SHOULD do something.
There are very good reasons for developing an Alexa Skill (or a Google Action). But the evidence to date shows that worthwhile applications are unlikely to come out of the Marketing Department.
The Draw
Amazon’s Alexa platform is amazingly open. Anyone can build a Skill (i.e., an application) and publish it to Amazon’s Skill marketplace — for free.
Not “everyone” can actually build a Skill, and it's not as easy as Amazon claims — for one, you have to navigate Amazon’s Byzantine Lambda platform. But, as of January 2018, over 24,000 Skills had been published. Although most of them were created by hobbyists, many are useful and engaging.
Amazon is not alone. The Google Actions platform launched last year, making similar app development possible for Google Home, Android devices and iOS as well. Apple, Microsoft, Samsung, and a host of Chinese companies are getting in the game too.
The related hardware numbers are ascending rapidly as well. By 2021, research firm Ovum predicts that intelligent assistants will outnumber human beings.
Source: Ovum
Why Brand Marketers Are Failing
Brand marketers have jumped at the chance to do something with the platform, but have struggled mightily. To this point, they are almost all failures: Sloppy user experience, poor branding, and limited usefulness abound. Worst of all is a consistent lack of authenticity.
Why have so many brands failed? One simple fact: they are approaching Alexa with too much advertising baggage. Rather than take the opportunity to add something useful and unique, brands have consistently shown that they don’t understand Alexa’s potential. And they don’t treat it like the new medium that it is, instead seeking to fill it with ephemeral novelty.
Alexa is not an advertising platform — it is an information delivery vehicle.
If you want people to use your skill, don’t treat it like a commercial or a social media post. If I ask for a recipe, give me the recipe, but don’t try to sell me Campbells’ soup. If I want to learn about how whiskey is made, tell me the whole story, not just Johnny Walkers’ history.
Some might argue that even if these efforts don’t work great, they’re worth trying. After all, they get people talking about — and with — your brand. But is it worth it? In most cases, no. Clumsy gimmicks reflect poorly on your brand and get in the way of more useful Skills your customers should actually be using. Your customers deserve better than a “rapid fail” development cycle — don’t treat them like beta testers.
Rules for Brand Building on Alexa
Despite the lackluster results so far, there is hope. Brands can and should explore the potential in Alexa, provided they follow four common sense rules:
Be careful - and don’t try to duplicate your mobile app.
One of the most common complaints is that Alexa Skills don’t work as well as the parallel mobile app. So before building a duplicate feature set, brands should ensure that their Skill works as well or better than the mobile version.
Sample fails: Uber, Lyft, Dominos, Fandango - Location matters, a lot. As does account linking. All of these Skills suffer from horrible ratings despite the success of existing mobile apps.
Sample Success: Capital One - Simple and effective, it delivers basic account information more rapidly than a mobile app. And to address security concerns, the Skill uses a four digit PIN to pull any information.
Be useful — and comprehensive.
Most Alexa users are looking for something. They want help. This offers a great opportunity to engage in a useful way. But if you limit your skill to the narrow confines of your particular brand story, you won’t see users coming back, and your investment will be left in the trash heap of unused Skills.
Sample fail: Johnny Walker - Despite the Skill’s 5 star rating, its pushy user interface fails on many fronts. Narrow knowledge base in “Whiskey 101” is particularly frustrating.
Sample success: Purina - Gives practical information on dog breeds in a friendly, easy to navigate interface.
Be authentic — and non-commercial.
Alexa users know you have something to sell, but you don’t need to sell it. It is simply too early in the evolution of the Alexa platform to be pushing users to a transaction. Lay off heavy-handed product promotion and you will be rewarded with repeat engagement. You will be seen as a resource rather than an item in a shopping cart.
Sample fail: Moet Hennessy - “Bottles and Bubbles” offers little useful information, described by one reviewer as “a long form commercial.”
Sample success: Tide - Once derided by reviewers for its hard product pitch, Tide has streamlined the experience and now offers to send directions via text or email.
Mixed results: Stubbs BBQ - Dripping with authenticity thanks to the voice of the Stubb himself, but the results are sometimes less than helpful: “When cooking fish, you need heat.”
Testing, testing, and more testing.
A great application experience can be ruined by buggy implementation. Be cautious and budget for ongoing development after launch.
Mixed results: Zyrtec - A great concept — personalized pollen forecast — but executed poorly. Requires multiple set up steps before delivering anything of value.
Sample fail: BMW - Plagued by bugs and missing key information. Not a smart move for a luxury brand.
Sample fail: Patron Tequila - Under the guise of age gating, requires an email address prior to use. There are better ways around this.
Context Matters
Last, don’t forget context of the user experience. Alexa users see their devices as tools. Brand marketers should be asking “What kind of tool do they need?” rather than “What kind of tool can I build?”
Further Reading
A nice application of product design principles to Alexa Skill development: “Nobody Cares About Your Amazon Skill”
“Google Set to Dominate Apple, Amazon, and Samsung in Race for the Best Voice Assistant”