BlogSkill Store Optimization

The Skill Store Optimization Stack

By November 9, 2020 November 11th, 2020 No Comments

I have had plenty of calls with users of our Skill Store Optimization (SSO) solution and quickly realized that users have a different level of knowledge about SEO (Search Engine Optimization) or ASO (App Store Optimization) and some could be a little lost about which levers to pull to gain more traffic for their Alexa skills. Remembering the famous ASO stack from our Berlin friends at Phiture, we built a framework illustrating how all the pieces of SSO are linked together.

The Skill Store Optimization Stack is a cheat sheet for those having launched (or about to launch) an Alexa skill and wanting to improve their skill store presence. In this post, I will go through the stack and share examples of tactics to leverage with a focus on the first two layers of the SSO Stack:

  1. Increase visibility: how to get your skill in front of more people? Be it on the Alexa app, Amazon website, or when users ask Alexa a question.
  2. Increase conversion: how to convince people to enable/activate your skill?

 

Skill Store Optimization Stack

 

1. Increase the visibility of your Alexa skill

The first goal of SSO is to increase your skill visibility. For this, you need to get in front of Alexa users, either searching for something specific or browsing the store, trying to discover new skills. There are various ways to achieve this.

Skill’s Assets and Keyword Optimization

Alexa users looking for something specific will either search on the Alexa app, the Amazon website or ask Alexa. While the first two cases (Alexa app and website) have lots in common with SEO and ASO, the latter is more complicated. Alexa actually needs to understand your intent (skill recommendation) and not serve a generic Wikipedia result. But in essence, the three cases are similar.

At Voxalyze, we estimate that two-thirds of skill activations come from search, regardless of the surface (Alexa app, website, asking Alexa directly). So, what are people searching for? This could be generic keywords (“meditation”), sometimes with typos (“meditaiton”) or branded ones (“New York Times”). What Alexa skills marketers need to keep in mind is that search is at the bottom of the conversion funnel: users already know what they are looking for.

The goal with SSO is ultimately to acquire customers that you will be able to monetize. One of the main levers and a massively untackled opportunity is keyword optimization. This is a great source of high-quality users – provided your target and rank for relevant keywords – as keyword optimization will put your Alexa skill in front of users with specific intent. Here you want your skill to rank on keywords with high search volume and relevance to your Alexa skill. Given the limited inventory of voice, be it on the Alexa mobile app or when Alexa answers a question, you should focus on getting your keywords in the top 3 positions.

Let’s look at how to do this. Put simply, you need to:

  1. Identify the high volume and relevant keywords, search terms, and questions that users will ask Alexa for or type in the Alexa app’s search bar. Ideally, you want to identify those with a low level of competition. 
  2. Integrate those keywords in the assets of your Alexa skill: invocation name (if your skill isn’t live yet – you can’t change it later), public name, keyword set, one-sentence description, and detailed description. The presence of a given keyword in those different places doesn’t have the same weight. A keyword in the public name or the one-sentence description will have more weight than in the long description.
  3. Analyze the impact of the changes and iterate. Keyword optimization is a continuous process. To continuously rank high on important keywords, you need to keep optimizing as search trends, the Alexa ranking algorithm, or simply your competitors on given keywords keep evolving.

Here is a snapshot of one of the skills we use for testing. It ranks #1 for brain exercise (great), #2 for brain training (great), and #6 for brain game (not so great).

Keyword optimization is a fantastic opportunity but also a complex topic. I will be diving deeper into it in an upcoming series of blog posts.

Getting featured

That’s every Alexa skill publisher’s dream: getting featured. For some developers, this has been and is still their number one source of traffic. The best way to achieve this – provided you have an awesome product, of course – is to build relations with representatives of Alexa in the countries you are interested in and “chip in” some favors. Those could be things like incorporating a new Alexa feature into the latest release of your skill, a promotional event (Halloween, for example), or simply allowing them to use your brand when communicating with either the users or developer community.

Localization

Localization brings your Alexa skills to different locales, where the same language is spoken. So, for example, you could localize a US skill to the UK, Australia…etc. Localization is much more than simply translation. As a developer, you need to understand the context and identify the relevant keywords and search terms. A classic example is the keyword “football”, which would be translated to “soccer” in the US. So a skill about “football scores” would mean two very different things in those two countries.

Of course, not all Alexa skills can be localized, as is the case for local restaurant skills.

Cross-linking of Alexa skills

Cross-linking skills is the process of linking skills together. Practically, within the experience of a given skill, usually at the end,  a second skill is mentioned. This would typically be something like “as you have enjoyed this skill, you may want to try this other skill.” While interesting on paper, there are certain limitations, namely:

  • You need to own the two skills. So, this isn’t for brands or single-skill developers.
  • The two skills need to have an audience with similar interest to have users follow the recommendation (hard to link “vegan cooking” with “burger recipes,” for example)
  • You can’t track the results, making it hard to optimize

If you are a large skill studio, it may still be worth setting this up, though.

Ratings & Reviews

Ratings & reviews play a role both in terms of increasing visibility as well as increasing conversion for your Alexa skill. Let’s cover here the visibility part. If we disregard the fact that Amazon uses reviews as a source of metadata for its search algorithm (minor impact), reviews are mostly used as a proxy for skill quality.

The Alexa ranking/search algorithm requires data on the perceived quality of an app. Retention and engagement metrics are great indicators for this, but this doesn’t apply to skills for non-everyday-use (book an appointment at a car repair shop, for example). Amazon, therefore, uses ratings & reviews as a proxy for skill quality. So, the better the ratings and the higher the number of reviews, the more visible your skill will be. 

 

2. Increase the conversion of your Alexa skill

Increasing the conversion of your Alexa skill is increasing the rate at which people who discover your skill (via the app, the website, or an Alexa enabled device) convert into enablement. It requires a mix of creativity and science. To be successful at Skill Store Optimization, developers need to increase both the visibility of their Alexa skill and its conversion. To take a financial analogy, visibility is revenue, and conversion is the gross margin percentage; improvement of  both is needed to have a massive impact on your bottom line (enablements)

Skill Assets

Once a user gets your Alexa skill suggested, there are significant variations depending on the surface used. Let’s look at each of them and how they differ.

Alexa mobile App

The Alexa mobile app presents a skill in two steps. The first one is via the ranking for a given search query, where the name, category, launch phrase, and ratings appear. Then, after clicking on the desired skill, the user will see the beginning of the skill’s detailed description, the “what’s new” section, launch phrases, permissions, links to the privacy policy and terms of use, In-Skill-Purchase (ISP) information.

Alexa Website

When searching for an Alexa skill on the Amazon website, the user also goes through a two-step process. The first step is again presenting search results with name, ratings, and one-sentence description being displayed. After clicking on the selected skill, the user access the full skill page with name, launch phrases, detailed description, permissions, links to the privacy policy and terms of use, and ISP options being displayed.

Alexa-enabled device

When a user asks Alexa for a certain task, Alexa suggests some skills but hardly displays any information: only the skill name and the ratings. 

The key takeaway here from a skill assets point of view regarding conversion is the importance of the name, one-sentence description, and the beginning of the long description.

Ratings & Reviews

Ratings & reviews play a role in increasing the visibility of your Alexa skill and increasing the conversion to enablement. 

Apart from an increase in visibility, the number of reviews and skill ratings significantly impact conversion. Amazon lets users give a 1-5 stars rating for any given Alexa skill. The user could either do it on the Alexa mobile app or Amazon website and then write a review. Let’s face it, the process is cumbersome, and hardly any user will go that far. Alternatively, at the end of a session, the users sometimes get asked to give a rating. You can’t trigger this, unfortunately: it is done by Alexa’s own algorithms.

This is a virtuous circle: the more positive ratings you get, the more visibility our app gets, the more conversion, and ultimately the more users….who then generate more positive ratings. The challenge is to get to the point where your skill is in the circle.

Several studies in the mobile ecosystem have shown the importance of ratings, and so far, there is no reason to believe that this is different in the Alexa ecosystem. You will find below the one of Apptentive – a mobile customer experience solution – where the numbers speak for themselves.

 

 

You can trigger rating requests in the mobile world, and it’s common practice to trigger those in positive setups (happy users). It isn’t possible – officially at least –  do this on the Alexa ecosystem. Another noticeable difference between the two is that ratings and review remain available even after several updates of your Alexa skill, while in mobile, those are reset after each update. This also means that a buggy or sub-optimal experience once could have a lasting effect on your ratings.

Icon

The icon of your Alexa skill is visible all the time to the users, except on screenless devices, obviously. This makes it a key element to optimize for. The screenshot below that I took a few years back at a conference comes from the mobile world and shows how much of an impact changing an icon could have. 

 

There are plenty of tips on creating a great mobile app icon, and most of them do actually apply to Alexa’s skills. Here are the most common ones:

  • Make the icon clear and simple.
  • Stick to your brand guidelines (if any)
  • Add borders to the icon.
  • Don’t use text (think about a user checking Alexa skills on his smartphone!)
  • Check the colors used for the industry or space you are in

Alexa’s challenge is that you can’t A/B test, which makes it super difficult to optimize. The mobile world has tools to do so, the Alexa world not yet (maybe something we at Voxalyze should be looking at). In the meantime, you could use panels or mockups.

Launch Phrases

Launch phrases are the most visible asset on the Alexa mobile app. The common practice is to include the skill’s name into the launch phrase, especially if your skill name includes your brand name. There are some alternatives, though, and many ways for developers to be creative.

 

Name Free Interaction

Amazon recently introduced the Name-Free Interaction (NFI) to certain locales and selected Alexa skills developers. What happens with NFI is that when a user speaks to Alexa with a requested action that does not include a skill name, Alexa recognizes the intent and selects a skill to fulfill the request. Practically this means that a user doesn’t have to memorize a skill’s name, just a keyword.  There are obviously different candidate skills that Alexa considers for a given keyword before picking the one that will fulfill the user’s needs.  There are three things you should remember:

  • For your Alexa skill to be considered, you need to make sure it uses the NFI Toolkit recently made available by Amazon. 
  • You could specify the keywords and phrases to be considered by NFI.
  • The ranking algorithm is the same as when people are searching for skills based on certain keywords. 

So, NFI is more or less a simple booster of conversion, but you still need to make sure you rank #1 on those high volume / high relevance keywords. 

 

3. Black Hat SSO

Black hat SSO, just like black hat SEO, is a practice aiming at increasing the visibility of your skill (ranking higher in search results), which is against Amazon Guidelines. The problem is that those guidelines are kept necessarily vague as the technology and ecosystem evolve, and it is therefore not easy to know when you are crossing the line. What is clearly within the guidelines is named white hat, what is clearly against black hat, and the big blurry space in between grey hat SSO. Some of the tactics we are aware of in the Alexa world are:

  • Incentivized or paid reviews & ratings – Black
  • Own reviews & ratings – Grey, but everyone does it!
  • Negative reviews on competition – Dark grey
  • Bot install/enablement and retention – Black
  • Keyword stuffing – Grey
  • Using fake brand names – Black
  • Cross-linking of skills via dynamic content – Black
  • Asking for review via dynamic content – Black

I’m sure there are plenty more, and that new hacks will appear at some point. If you need some guidance, keep in mind that the common practice is to stay within the white and light grey areas while keeping an eye on what is happening on the dark side.

 

4. Skill Store Optimization Tools

I have described above the various levers to increase both visibility and conversion. SSO, like every optimization process, requires a lot of information that can be tedious or even impossible to gather on your own, and that’s what we at Voxalyze have decided to tackle. Our vision is that our solution will be the ultimate go-to place for all your Alexa skills’ marketing needs. I won’t use the famous Amazon “Day One” analogy, but let’s say that we are still in the early days. So far, we cover the following:

  • Keyword rank tracking: know for which keyword your skill ranks and at which position.
  • Ratings tracking: see how your ratings evolve.

These features will be released in the coming weeks and months:

  • Keyword research: find out which keywords could be great opportunities for your skill.
  • Reviews tracking and sentiment analysis: what do users write about your skill.

So, why don’t you sign-up today, try it out (it’s free), and see how you can boost the presence of your skills in the Skill Store.  If you have any suggestions, comments, or questions about the Skill Store Optimization Stack, please let me know!

Alexis Hue

Alexis Hue

Alexis is Voxalyze cofounder. He is passionate about the audio space and loves using data to unlock and enable growth.