We’re getting to the stage where mobile internet usage is set to overtake desktop. In some parts of the world people’s first experience of the internet is through their low-cost smartphone. This new way of consuming content is seeing a shift in the way in which companies like Canva work and the ways in which we seek to reach millions of users across the globe.
While the reality is that everyone in the company is working on ‘growth’, the dedicated Mobile Growth team at Canva have their own focused set of goals including to:
- Help users find our mobile apps.
- Once they’ve got the apps on their mobile, help users get the most out of the product from the first time they use it.
- Get them back into the app if they have churned.
‘Design anything. Publish anywhere’ is our mantra. It’s through this mantra that we enable our users to use Canva on their preferred choice of device, through which we then hope to build some of the most loyal users in the world. With the increasing importance of mobile, we’re always looking for ways to increase the number of our cross-platform users and ensure we’re always offering them a smooth experience.
The mobile growth squad at Canva formed late last year with the purpose of testing everything from: landing page CRO, driving owned media installs through desktop placements, email and push campaigns, App Store Optimization, to in-app experiments. We work in an environment where we are constantly testing hypotheses and aren’t afraid to change things up when we see that they aren’t working as successfully as we would like them to. It’s important for us to learn as fast as possible and understand that solutions to all problems can’t be solved in one iteration.
In this article I’ll talk about a few of the things the Mobile Growth team at Canva have worked on over the past 12 months, and some lessons learned along the way. I’ll group these learnings into three basic areas: Experimentation, Internationalization, and App Store Optimization (ASO).
Some of Canva’s Mobile Team
As with most experiments or product growth initiatives at Canva, mobile growth experiments themselves are focused on getting users to the app, getting them to publish their first design, and driving them back into the app. Here are a couple of examples of experimentation around owned media acquisition and retaining users through push.
Experiment 1: Growing the cross-platform user database
With 10+ million users already designing in Canva, we launched a dialog box on desktop that we show to our users after they publish a design. This meant that after they had published on desktop we were able to promote our mobile apps to them.
Why is this useful?
Once users have reached their ‘aha moment’ on the platform they found you on, in this case desktop; we can take the opportunity to let them know about the cool things they can do on their mobiles.
Prior to the Android launch we used it as a way to collect phone numbers and these users received an SMS once we launched.
Sending users directly to the app stores on desktop (a) breaks the user flow and takes them away from your product and (b) isn’t a great user experience for desktop users—while it’s a little better on Google Play, you can’t actually install an iOS app from desktop. We wanted to make it as simple as possible, by allowing users to SMS themselves a link to their mobile. They tap on the SMS and are taken to the relevant app store. This is fairly common practice across many websites but we ran an experiment to drive more app users from this placement.
The original SMS dialog was two steps—first you selected iPhone or Android, moved to a different screen and then we asked you to enter your phone number. Users were required to enter their phone number with the correct country code, otherwise the SMS would fail.
The experiment involved combining the two steps of phone number entry into one; and also detecting the user’s country and pre filling their country code for them so they could enter their phone number more naturally.
We were confident that the variant would perform better than control as it was quicker, easier and less prone to human error. We decided to test out our messaging for the two variants to get some further learnings from the test.
V1: We showed a mobile device—pretty simple right? We’re pushing the mobile app and here’s a giant image of a phone you can’t miss.
V2: And we created a text-heavier variant where we actually listed out the benefits of the mobile app. But we assume people don’t always take the time to read things right?…
The winner of V1 and V2: V2 saw 21% more phone numbers submitted over V1. It turns out people do read things, and in a world where we’re continually told to “Go mobile”, it can help to lay out specific benefits and uses of the mobile app for your audience—test and see if it works in different situations though.
Winner overall: As was expected, reducing the number of steps and helping users enter their phone number in a more familiar format more than doubled the average number of phone numbers submitted each day.
Experiment 2: Push notifications
Push notifications are a great great way for us to engage our mobile users. At Canva we use them for activation and retention purposes. While we send many types of push notifications; many of our engagement and retention pushes are based on current events.
Push is a great channel for this kind of thing for a couple of reasons:
- It’s quicker to setup than an email.Push are relatively quick to set up for events like Valentine’s Day, Cinco de Mayo, or Mother’s Day.
- The copy is under 200 characters—including the title and body—so we don’t have to spend time writing subjects, pre-headers, and the body copy for an email.
- Setup takes less than 10 minutes, there’s no coding involved and there is less testing involved as compared with an email.
- Some days are mobile days (so let’s capitalize on that!) We see events and days during the year where mobile usage spikes and desktop does not. There are certain use cases for occurances and days like these, such as sending Valentine’s Day cards or posting Instagram stories, where push notification are better suited than an sending email.
An example of push notifications
TIP: Set a control group
If you send a push notification on Christmas Day to your entire user base and are pleased to look at your results and see a spike in Christmas cards sent, you may just be looking at the impact of the day—not your push alone. The point of sending these campaigns is to bring about an increase in engagement above and beyond what your users would have done regardless of your push.
For event-based push notifications we see improvements of 15% in designs shared in the group that received the push over those the didn’t. Both groups saw a spike on the day of the event because people were more likely to share a design on the day, but we see that the push caused an additional uplift.
Push notifications allow us to test emoji vs. no emoji, shorter copy vs. longer copy, day before event vs. day of event, and several other such factors. However, at first our event based experiments began by purely testing if our push themselves were having an impact on these already high-volume days at all.
Experiment 3: Improving the cross-platform experience
With growing numbers of people opening their email on mobile, we wanted to improve the engagement and acquisition firepower of our email campaigns. We use Branch, a mobile linking platform which powers deep linking. This allows us to link directly from our landing pages, emails, and other assets to our mobile apps.
Where the optimal experience on mobile is within the mobile app—editing a template for example—users that click on the CTA in an email will experience one of two things:
Use case #1: User already has the app installed and taps on a design in an email from their mobile device. They’re taken straight through to edit that design.
Use case #2: User flow for someone that does not already have the app installed. After installing the app and signing in, they will be deeplinked to the template they tapped on in the email.
If the experience is better on web, e.g. an article on our Learn Blog, we take users directly to the web URL and don’t thrust the app on them. However, where there are templates highlighted in the article, we use Branch links to deep-link mobile users into the apps to edit.
Improving our cross-platform experience is something my team and I are continuing to test and improve on.
Canva’s mission is to empower the world to design. With over 80% of mobile users in Brazil and India using Android devices, launching the Canva app for Android in November 2017 had a huge impact on growth at Canva. In a few short months we quickly reached five million Android installs and international markets such as these will continue to be the key to our overall mobile Monthly Active User (MAU) growth.
Canva’s mobile apps are localized into more than 50 locales, but there are a variety of factors we need to take into account when planning out our mobile growth in international markets. Here are a few differences in markets beyond just language to think about:
Device memory or storage as well as network speeds vary across the world. These can affect retention rates or conversion from the storefronts if a user is simply unable to install the app in the first place or it’s impossibly slow to use.
We look to optimize the experience from the moment users open the app, sign up, find a template and then share their design. We know that a one-size fits all approach is not going to provide us with maximum growth. We also see that push notifications or in-app experiments can vary between markets, so this is something that requires testing.
Sensitivity to price and localising prices.
We know that there are variances in price sensitivity and willingness to pay across different markets and platforms across the globe. This could have an impact on activation and retention rates for your app—both stores make it fairly simple to localize pricing for purchases.
App Store Optimization.
I’ll dive into this in more detail later but going back to the point about differences beyond language: You can complete localized keyword research to optimise your titles, descriptions and show localized screenshots, but you need to optimize for conversion on a market-by-market level. Google Play allows developers to A/B test in all languages your storefront is available in and we have seen different results for the same experiment in different markets—even among the US, UK, and Australia. Additionally, there are opportunities to give your Storefront a truly local feel and it can be worth putting a little extra time in.
Mobile Operating System Market Share Brazil: May 2017 - May 2018. >80% Android. Source.
Mobile Operating System Market Share India: May 2017 - May 2018. >80% Android. Source.
App Store Optimization
There are no two ways about it, you need to put time into App Store Optimization (ASO) and it’s something I personally love working on. With over 60% of installs coming via App Store search the Storefronts are a vital piece of the getting-your-app-discovered puzzle. The groundwork needs to be laid to set up a good foundation for organic search and for optimising conversion when you do start sending more people to the Stores from other channels like landing pages, paid or owned media.
Both the App Store and Google Play have made some significant changes in the past year and its important for us to look at ways to work with the Stores and improve our product pages to make the most of these changes.
- With the new Google Play Storefront removing the feature graphic and bringing your screenshots further up the page there may be some wins with screenshot experiments (again, think local); and testing video/no video on your store listing may deserve revisiting.
- The Google Play Console will soon give developers access to store visitors, installers, install rate and retention rates by keyword; the ASO world rejoiced when this news broke. It’s not reported by country (fingers crossed this will happen soon), but here’s an example of how you could use this data: Identify keywords that are driving high volumes of visitors, with a poor conversion rate. You may be ranking well for keywords that aren’t coming across to visitors as relevant to your app. If they are relevant, try referencing the keywords in your first one or two screenshots to drive an increase in conversion rate for keywords you’re already ranking well for. If they’re not actually relevant, you can focus your metadata efforts on ranking better for high converting keywords.
The new Google Play Store design puts more emphasis on screenshots. Source.
The App Store
With the massive overhaul of the App Store in 2017, we saw not only a thing of beauty, but more than just lists of apps shown in search results and curated lists. The coveted App Store feature which always brought about a spike in installs from users that weren’t really looking for your app—those who don’t have the greatest need to use your app and therefore are less engaged—now lives on in search results in the form of Stories. You get a longer life out of the editorial piece and are bringing in users with higher intent. In saying that, where you previously had five apps in the top five of search results—below Apple Search Ads in many markets—your app is often now competing with a Story and promoted in-app purchases within search results. Making top rankings across high volume keywords all the more important.
On the new App Store apps, Stories written by Apple Editorial, and promoted in-app purchases are all shown in search results. Prior to this, you were only competing with other apps for top position and visibility in search results. Now with so much more going on in search there are more opportunities for acquisition and driving revenue from search.
At Canva our cross-platform users are our most loyal, so growing mobile and improving their experience is growing everything. It pays to come at it from all angles. Not just paid or organic, not just ASO, not limiting your changes to the in-app experience. We’ll continue to keep testing, so watch this space.