This post will address two fundamental questions: One, what makes a mobile app “great”? And, two, how does one distribute that app to make it greater? (I’ll answer this second question next week, below.) Over the past nine months, I have been pitched by at least 200 different teams to evaluate their mobile app, most of them while I was working out of Javelin Venture Partners — and I have been working very closely with a small handful of pre-launch mobile teams on the operating side. I have devoted all of this time to addressing these questions and trying to develop a thesis, framework, and playbook for entrepreneurs and investors alike. As with any thesis, I make assumptions — such as, I’m evaluating iOS apps first, though there are some caveats emerging. I will try to cite them along the way, so please bear with me. This post, then, is the result of that work, and as with most fast-moving things, I will continue to update this as I learn new things.
Part 1 of 2: INGREDIENTS OF A GREAT MOBILE APP
What makes a great mobile application? Here are the main ingredients. To be “great,” an app must have many of these characteristics:
- Strong Daily or Weekly Active Use Case: Is there a reason consumers can use or benefit from your app every day, or at least, a few times a week? In the same way we view measure web activity and retention, the mental model around a great app is one where the product provides the consumer with a reason to open the app at least a few times a week and be immersed in the experience. (For apps that promise to work in the background, push notifications for reengagement and reporting become more important, though usually are not considered “strong” use cases yet.)
- Strong Models for Attention and Retention: When a consumer opens your app, can you command their attention for a meaningful amount of time (important for ad-driven products), and/or can you provide them with utility that provides demonstrable value? For years now, the masses mock properties like Instagram or SnapChat because they don’t know how these services would ever make money. But, for apps that harness consumer attention (measured in engagement time per use), those seconds and minutes add up. Second, how often do people come back to the app?
- App Categorization Bias: Not all app categories are created equal. Apps that are games, or communications-driven — not just 1:1, but 1-to-many and/or broadcast, in order to harness network effects –or deliver news or information, or are tied to media like the (video) camera or sensors (like GPS, etc.) statistically stand a better chance of reaching mass consumer scale, in terms of downloads and attention. Similarly, apps that create a mobile marketplace (Uber) also stand a much better chance of creating value because, just as they do online, mobile marketplaces that can quickly achieve liquidity earn great advantages.
- Clear Mental Models: Does your app do one thing, and do that one thing well? There’s an app called “Twist” which only crawls your phone’s calendar and address book to automatically send SMS updates to those folks who are going to meet (in the background), to simply automate the “I’m on my way” text messages by using background location services. The easiest way to hook in a new user is to provide one single point of utility (the mental model) and then slowly expand from that. The easiest way to dissuade a new user from reengaging with an app is to throw an entire web product’s functionality inside the app. This happens way too often and we all know the results.
- Lean-Back Experience or Background Service: Most consumers want to open an app and have it do one thing or consume information/media. Asking them to do lots of work, like going through elaborate menus or creation-steps, is simply unrealistic. Most mobile audiences are casually browsing their apps, and very few — like Twitter, and Facebook, for instance — get their audiences to continuously engage with their media. Most people want apps like Pandora, where you open the app, register, and somehow — automagically — get what you want. Separately, if an app can require simple registration and then run in the background and engage based on time of day and/or location (like Highlight), that’s good too, although I still wonder about consumer’s willingness to allow the GPS sensor to run given the battery drain.
- Product Virality: How does an app spread to others? Outside of the elusive “word-of-mouth” virality, all successful apps need some growth model to reach “venture-scale,” if that is the team’s goal. When asking for venture investment, the app simply needs to be designed in a way that it can grow by its own engines. This could be through social invitations and interactions, or through gaming or multi-channel communications. Conversely, apps that don’t provide models like “single-player experience” or “one-to-many” communications face a tougher road to growth. (We’ll talk more about why this is important below.)
Once you’ve built and/or identified an app that has great ingredients, the work is not done. In fact, it is only just beginning. What I’ll describe below is mostly greenfield and evolving, and I suspect it will change on a quarterly basis as more and more people are forced to experiment — and innovate — around mobile app distribution. In this section of the post, I’ve attempted to lay out all the steps a team must consider in order to increase their chances to win and dominate at distribution. Even great apps that meet the criteria listed above could plateau unless they’re given every opportunity to be distributed at scale. Here are some tactics I’ve identified that mobile app teams should consider at least six (6) months prior to launch, with the obvious caveats that not each category will apply to each company. [I’ve based this list and checklist on a framework I used while in venture to help evaluate mobile apps and their teams’ ability to execute on a plan like this.]
Using the Web to Hack Mobile Distribution
Up to the middle 2012, the phrase “mobile first” dominated our consciousness, but it lacked a key ingredient: the Web. Yes, the Web. One thing I learned from a handful of startups who were able to raise significant Series A rounds and have a mobile product was that they consciously leveraged the Web as part of a broader strategy to give themselves more ammunition in the fight against distribution challenges. [For more context and ideas, watch this conversation I taped with Rishi Mandal, the founder of Sosh, and read this postthat focuses on iOS distribution in particular.]
Typically, the lead mobile teams will be in charge of initial user testing. The key for alpha testing is to get the app in the hands of the types of people who can give you honest, unfiltered feedback from a variety of viewpoints. More specifically, you’ll want regular everyday users to test the app onboarding, flows, and notifications, as well as those who are designers, investors, potential partners, and the like. It’s a good idea to make a list of the important constituents for your app’s use cases and then map back to a list. In general, it’s best to make sure that those invited to Alpha don’t publicize your efforts and also agree to giving you feedback. Simply having alpha testers without feedback from each one is a waste. In addition to verbal/written feedback, the team will want to make sure its metrics gathering is set as well, to make sure who is really using the app and other patterns that may emerge from the data. [During and after the Alpha Test period, product and engineering will have a lot of work to do, so make sure to build in time in the schedule for this. Apple especially doesn’t make this any easier.]
I’ve seen some clever hacks around expanding Beta Tests recently. The beta should be a bigger, wider population of friends and insiders who can try v2 of the app and give feedback. But, in addition to individual feedback, the beta test is about reach and finding insights from the aggregate data. These insights hopefully inform both product and engineering on the final tweaks needed before launch. Again, Apple makes it hard to invite hundreds of folks to test your app, because they want you to make it public in their story as soon as possible. As a result, startups will use a variety of tricks to get around this — they will sign up for “enterprise” accounts with Apple, or use Hockey as a distribution service, or even hand out builds piecemeal to get more initial users to fine-tune before launch. Apps can also be launched in Canada’s App Store, unavailable here in the States, and leverage services that find initial users to get even more data before launch.
The Elusive Word-of-Mouth
This is also the time in which startups can begin to build buzz for their app. The latest and greatest example of this is Mailbox, which slowly released beta versions around the Valley and then, upon launch, executed on a brilliant velvet-rope reservation system to further stoke demand. Of course, none of this works unless the app meets the requirements above, which Mailbox did. Generating buzz works well when a product is new, when the product invites usage frequently every day, and provides utility and/or delight. There’s additional angles that need to be tapped into, as well, such as a framing around the mental model of when users should use your app, as I discussed with Josh Elman in this conversation.
Launch Plan and Checklist
I’d encourage startups to begin collecting emails and phone numbers of potential users as soon as possible. This creates a distribution database, as well as a memory in the potential user’s mind. It also creates an opportunity for the company to communicate with users leading up to launch. Furthermore, a team can leverage traffic patterns on the web to divert traffic to a variety of landing and signup pages, with the benefit of tracking conversation — though it’s trickier to track conversion of the app download when your app is ready. As a hook, many startups will use incentives for folks to sign up (“save your spot in line”) and to share (“tweet this to move up in line”) and so forth. It may sound silly, but people actually do this and it can work.
I’ve taken you this far, and there is more to go. Below, I briefly list out the different categories you and your team will have to pay attention to at launch and beyond. Instead of listing all the tactics in prose, I’ve listed them out in bullet-list form and ask questions for you and your team to consider. This is the exciting part, because as Mailbox demonstrated, there are 101 unknown novel product marketing techniques to experiment with to breakout from the noise in the app ecosystem.
Goal: Tell the world about your app, and to be be aggressive.
- Press: Do we have active press relationships that we’ve kept in touch with, or do we need to consider a PR consultant or PR firm to help?
- Web marketing: Do we have a series of landers in place to buy or divert web traffic to our app download page and trace the efficacy and CAC of this approach?
- Mobile marketing: Do we want to consider advertising on mobile products themselves, and if so, what’s our budget for that?
- Reviews: How can we encourage thousands of people to positively review our app in the app store ratings?
- Promotions: Are there any direct or cross-promotions we can tie-in with launch to increase the initial pop?
- Friends: How do we make sure each and every one of our friends and former colleagues downloads and uses the app?
- Emails: What’s our philosophy and system for emailing users as an engagement tactic? How frequently will we email them and for what trigger?
- Notifications: (Same as email above, but for push notifications on the device itself.)
- Open Graph and Twitter: Do we believe the actions users take in our graph would dovetail either with Open Graph or being mentioned on Twitter?
Goal: To make sure the masses can find your app.
- SEO: Can we leverage SEO on the web to continue driving traffic to our app download pages?
- Mental Models and Branding: What are the top use cases for our app and how can those be most easily communicated to the average consumer?
- Friends’ and Colleagues’ Networks: What incentives can we put in place to get our first-order networks to share the app with their networks?
- Word of Mouth: How can we get people to talk about our app with each other?
- Public Relations: Do we need to engage with PR even after the launch?
Goal: Win at distribution.
- Email: How well is email working at driving new users and engagement in the app?
- Social: How well is our brand and product being leveraged by users on Twitter or Facebook?
- Platform Order: If we launched iOS first, how do we think about building for Android next or Web next, and why?
- Specific Android considerations: Given Android device fragmentation, if we go Android “Second,” what device do we want to optimize for and why?
- Specific Web considerations: How do we bring users back to the web in a normal workflow, and/or how do we think about building for web given our mobile interests?
- Partnerships: Are there any relationships or natural partnerships we can leverage to get maximum exposure for our app?
- Advertising: Do have a budget for and want to consider more traditional advertising routes to get our app out there?
Goal: To increase engagement through rapid initial feedback loops.
- Feature-Testing: What is our team’s approach to rapid A/B testing to improve in-app engagement metrics?
- Product Roadmap: How does our product roadmap coincide with features we believe, or the data tells us, to be engaging?
- Digesting Feedback: What’s our system and process for rapidly collecting, analyzing, and acting on feedback that will positively impact engagement?
- Metric Evaluation: What metrics are we tracking and why are they important to engagement?
- Efficacy of Notifications: What are the metrics around our email and push notifications?
Goal: To make users come back often, to build a habit around the app.
Tactical Questions, which assume engagement metrics are tracking well:
- Lock-in: What features or services in our app have the ability to lock-in users?
- Product usefulness: When is our product most useful to consumers?
- Product announcements: Based on our product roadmap, how can we build or improve features that lead to better engagement and retention?
Any thoughts or reactions to this? Within these pieces, there are lots of overlaps, lots of things to discuss, and I’m currently making a big list and will be writing more on this very soon. Your comments are appreciated.