Nobody cares about your Amazon Alexa Skill

So here’s how to make them.

So you want to build an Alexa skill for your brand.

You did your research, raising your eyebrows when you discovered that there are already 15,000 skills in the Amazon Alexa skill store. You’re not actually sure if many of your friends have an Echo device yet, but if they do, they’ve told you they only really use it to find the weather or to play Spotify. Great- a glorified speaker. You’ve seen some impressive brand names on the skills store, but your execs are still on the fence about whether voice technology is the next big thing. Alexa’s voice sounds too much like a robot and if they’re honest, they’d rather play around with a VR headset.

Before the big boss can invest money in building a voice app, they need to understand its value- will it create a hub of dedicated users that we can upsell to? A quick check on retention rates isn’t promising. According to recent statistics, the average retention rate for an Alexa skill after a week is around 3%. Looks like people just try it once and give up. Not looking great so far. You sip your overpriced coffee nervously.

But the excitement around voice is real. You saw three Alexa adverts on the way into work this morning. You do some more research- it reveals predictions of 120 million Echo devices being sold by 2020, with over 40% of UK households having one by early next year. So surely it makes sense to dip your toes in the voice water?

This one had you singing Rihanna for the rest of the day

Many discussions and post-it posting workshops later, you finally get the green light. Let’s get a skill out there.

The four stages of releasing an Alexa skill

  1. You’ve got the confirmation email back from Amazon. They’ve certified your skill and it’s now available. You watch with pride as your voice app gets released into the skills store abyss, to join the 15,000 already floating around in there. Time for a celebratory beer.

2. Now you’re hungrily watching the reviews start to tot up from enthusiastic Alexa fans. They love that your brand has decided to get into this market, proving to be ahead of the game and cooler than your competitors. A couple teething issues with some of the functionality but on the whole it’s doing a great job.

3. Reviews start to slow down and before long, stop altogether.

4. Paranoia starts to creep in. Are people even using the skill correctly? Do they know about that amazing feature that you came up with? Voice apps have no hierarchy- unlike mobile apps, there are no call to action buttons to let users know what exciting possibilities are lurking round the corner.

You’ve reluctantly realised that you’ve now become another skill that’s contributed to the less than 3% retention rate. And you thought you’d break the mould.

But here’s the real stab in the back- you’ve just come across another dreaded statistic that says that according to an April 2017 study from GfK, nearly half of Amazon Echo and Google Home users report using their devices “regularly” or “all of the time.”

So people who own Echos and Google Homes are using them. They’re just not using skill.

This might just be the end of the road. The execs conclude that it was good to test the waters, but maybe the market just isn’t ready and attention should be shifted to something more promising.

Where did you put that VR headset?

Why people aren’t using your skill

Where’s the hook?

Niy Eyal’s Hook Model

According to Nir Eyal, author of Hooked: How to Build Habit-Forming Products, hooks are “experiences designed to connect the user’s problem with the company’s product with enough frequency to form a habit.” Eyal’s famous Hook model comprises four key steps that the most successful companies use to create sticky products that users can’t get enough of. This is extremely relevant to Amazon Alexa skills. You’re starting to wish you’d read the book before you started this whole adventure.

If you’ve built a skill that a user would only use on a whim, you can never guarantee repeat usage- the novelty will wear off quickly.

1. Trigger

What is the trigger for your skill? Why are people going to use your skill and what mood are they in when they open it? The most effective emotional triggers are in fact negative; boredom leads us to watch inane YouTube videos and loneliness turns us to the arms of Facebook. Think about what emotional state you want your users to be in and how your skill can prove that it satisfies it. For example; uncertainty on what to eat for dinner might lead someone to open up a recipe skill. A good trigger will become an itch for your Alexa users that only your skill can relieve.

If you’ve built a skill that a user would only use on a whim, you can never guarantee repeat usage- the novelty will wear off quickly.

2. Action

This one should be a no-brainer. For the Hook to be effective, users need to take a simple and efficient step to receive their reward. Voice technology could not make this easier. They only have to talk to Alexa to get what they want.

But think about the ways and utterances that a user might use; often developers end up alienating users because they’ve not accounted for all the different ways they might speak to their digital assistant.

Alexa analytics tools are useful here to monitor what your users are actually asking. From that, you can see whether you’ve missed any utterances and add them in before it becomes too much of a problem.

3. Reward

The reward stage is where users get their relief from the “itch” of the trigger. For Alexa skills, this can be the trickiest step to perfect, as the value of the reward rests entirely on Alexa’s response.

There are two ways that this can go wrong:

a) Your skill doesn’t understand what’s been asked and answers wrong. There is nothing more frustrating/less rewarding than Alexa failing to handle your request. This problem should only shrink as Amazon improves its technology, but it’s always worth double checking your code just in case it’s not her, it’s you.

b) Your skill’s tone of voice/ personality is alienating. This is completely within your power to nail. What personality do you want the Alexa of your app to have (newsflash: she doesn’t have to sound like a mindless drone. Besides, even if you would’t book her for your next stand up comedy gig, native Alexa does have some quirky personality traits of her own). Make sure your brand personality comes across in your skill.

4. Investment

This is possibly the most important part of the model, as it requires the user to give something back to the product that will help to make sure they return to it. This step is easily overlooked and is what stops the Hook model from looping round and starting again, preventing your skill from becoming habit forming.

The Hook is only complete when the user invests something, passively or actively, back into the skill. This could be usage data such as location or weather, or allowing a user to actively add to the skill with their own personal content.

Think of Alexa as your friend. If you have a conversation with a good friend, you’d hope they knew something about you. This contextual awareness provides an interesting, banter-full conversation and is significantly better than the banal kind of small talk you normally have to endure with your exec in the lift.

In the same way, your skill needs to be the platform for a two way conversation, where the data provided by a user is used to create a more personalised experience in the future. Too many skills are built in a simple question/answer format with Alexa as a lifeless, slightly more useful magic 8 ball. If you’re creating a recipe skill, use the user’s location and weather to suggest recipes that compliment the weather or what’s in season. It’s these small touches that will delight and keep your users coming back.


If you’ve managed to think about each of the Hook model’s four stages when building the skill, there should already be some good reasons within the experience for the user to want to use the skill again. The more they use it over time, the more of a habit it will become. Of course, this should only get easier as voice technology adoption increases, but understanding and employing a good Hook should ensure that as and when your skill reaches people it has the best chance of sticking.

And then you can put the VR headset back down.

If you enjoyed this story, please recommend and share to help others find it! Feel free to leave a comment below.

The Mission publishes stories, videos, and podcasts that make smart people smarter. You can subscribe to get them here.

The Mission

A network of business & tech podcasts designed to accelerate learning. Selected as “Best of 2018” by Apple.

Jess Thornhill

Written by

Voice Designer & PM| Founder | Previously @ycombinator 17 w/ @OpearloVoice | Previously previously @AccentureDigi

The Mission

A network of business & tech podcasts designed to accelerate learning. Selected as “Best of 2018” by Apple.