The Atlassian Marketplace generates over $200 Million (yes, that’s not a typo) in sales per year, which is over 20% of Atlassian’s total revenue. From the “State of Salesforce” report by Bluewolf, 80% of all Salesforce customers use at least one app from the Salesforce Store.
The revenue and engagement benefits are clear but creating a successful app store platform isn’t straightforward. Both of those companies built their own solution in-house with endless resources over many years. They had dedicated teams to design, build, maintain, scale, promote and manage their platform.
When you decide the time is right to build an app store and platform strategy, you should understand the pros and cons of building internally versus using a specialized provider. We’re going to break down the costs, risks and things to consider when making that decision.
Our Obvious Bias
Of course we believe that using a specialized provider like OpenChannel can be of huge benefit. We focus on supporting a very specific strategy but that strategy isn’t necessarily a fit for every company.
Our goal here is to help you understand what’s involved once you’ve decided to build an app store platform around your company and your APIs.
The Engineering Iceberg
App stores should be designed with simplicity in mind, guiding users down the right path and avoiding confusion. Reducing barriers means that complexity must be hidden beneath the surface. When we speak with engineering and product teams, they believe that’s a simple task and we often hear things like:
“All we need is a form for developers to submit apps”
If apps aren’t versioned properly, you won’t be able to track or moderate any changes made by developers.
“Let’s charge users with our billing system then transfer the money to developers”
You’ll become the merchant of record for the developer’s products, where any fraud or disputes are on your record. All of your transactions can be blocked if your fraud incident rates become too high.
“API integration is just a simple OAuth flow”
The user will also need to be able to install, uninstall and reinstall apps easily, and the app will need to retrieve the user identity, plan purchased, subscription status and API key.
High Abandonment Rate
As a result, many in-house attempts to build an app store projects can be abandoned because of unforeseen delays. Engineering teams become flooded with requests as complexity and issues rise to the surface. As higher priority development tasks for the core product pile up, the app store project ends up being abandoned
Complaints from Administrators
The in-house app stores that do get completed often have an unsatisfactory result. Administrators often lack even the basic features like reporting and analytics needed to make their platform successful and can resort to moderating apps with spreadsheets. This means a higher administrative cost and longer wait times for developers to get their apps approved.
Delays and Limited Success
Those that are completed also often launch up to 12 months behind schedule, while others have required a complete rewrite of the beta. Even scarier, in our experience more than half of companies never receive additional internal resources needed to create a second version.
Starting is Easy, Maintaining is Hard.
When making the build versus buy decision, it’s important to ask some of the following questions:
- Can I take my engineers away from building core product?
- What are the actual and opportunity costs of our engineering time?
- How will I maintain the system in the long-run? Will I need to have a team full-time to manage the platform?
- What infrastructure will I need to ensure reliability?
- Will it be enterprise grade (speed, up time, scale)?
- How long will it take to design, build, test, and integrate the system into our existing product and workflow process?
The Actual Cost
Using an average team size of 3.7, and an average fully realized engineer cost of $107,000 /yr, the cost breakdown starts to look like this:
V 0.1 – Streamlining app submission and management for third party developers
- Expected time invested: 3 months
- Expected cost: $99,000
V 0.2 – App and ecosystem moderation tools for administrators
- Expected time invested: 4 months
- Expected cost: $132,000
V 0.3 – App integration SDK, key exchange and permission flow
- Expected time invested: 6 months
- Expected cost: $198,000
V 0.4 – Collecting payments and paying out to developers
- Expected time invested: 8 months
- Expected cost: $264,000
V 1.0 – RELEASE TOTAL
- Expected time invested: 21 months
- Expected cost: $693,000
…And at every stage you’ll need to do a full cycle of design, planning, building, testing and integration.
That doesn’t take into account any ongoing maintenance and consistent platform improvements. These are arguably the most important aspects for long term success of an app store platform. The platform needs to support the many edge cases that will come up but that isn’t your engineering team’s core focus.
Engineering teams often stop actively supporting the app store. They are flooded with request for reports, statistics or manual interventions that result in a high ongoing maintenance cost and a poorly managed case of feature creep.
But That’s Understandable
App store platforms are relatively new. It can be difficult to define the scope of a solution across not just engineering but marketing, partnerships, and other product groups.
“When evaluating whether to buy or build, it’s critical to thoroughly understand total costs during the software life cycle — typically seven or eight years. This step is important, because 70 percent of software costs occur after implementation. A rigorous lifecycle analysis that realistically estimates ongoing maintenance by in-house developers often tips the balance in favor of buying.”
– Mark Lutchen, former Global CIO of PricewaterhouseCoopers,
To get a detailed breakdown of milestones with requirements, users stories and build estimates, you can use the OpenChannel online requirements generator.
The costs, risks and other considerations are not surprising but in order to use any specialized provider, there has to be very explicit benefits.
An app store involves collaboration among many teams to onboard, support and promote the platform and app developers. Using OpenChannel frees up your team to focus on making the app store and developers a success rather than having to design and build an app store and infrastructure. Your app store is only as good as the apps within it and we help you focus on what matters most.
“If we could have focused from the get-go on the actual partnerships and building programs for our ecosystem, instead of having to build our own version of the app store across all the platforms we supported, we would have been able to grow Box OneCloud twice as fast.”
– Indy Sen (Box, Salesforce, Mulesoft, Google)
Reducing time to market is an important benefit of any pre-existing solution. OpenChannel will reduce your time to market by up to 90%. It will save your engineering team months of development but there are other savings that matter arguably even more.
Being top of mind with developers and providing them a path to get their apps in front of users is important. If you’re in a highly competitive space (and who’s not) competitors may be looking to build an app store strategy or already have their app store available to developers and users.
The months or more to build an app store may result in years of work trying to recapture developers. They may have already built on other platforms and in some cases it may be to late. This allows your company to be relevant in the minds of developers faster.
Our administration tools also help you save ongoing time when tracking, reviewing and moderating your apps. Your app store is a product to be maintained and improved, not a project to be completed. The ongoing time saved becomes more valuable as your platform grows.
Service disruptions on your app stores not only affect the experience for your users but your reputation with developers as well. OpenChannel’s service is built on geographically redundant infrastructure, trusted by NBC SportsEngine with their 35 Million monthly active users and Jamf for their ecosystem of over 9 million apple devices. If one OpenChannel datacenter is compromised, load balancers redirect traffic to the other datacenters automatically with only seconds of down time.
Companies spend significant time and resources promoting their app store platform. When a marketing campaign takes off, you need to know that value won’t be lost. OpenChannel’s auto scaling and auto balancing infrastructure spans across multiple data centers to absorb traffic spikes and ensure a consistent user experience.
Our dedicated support team work around the clock to monitor and fix issues before they ever impact users and developers. OpenChannel is constantly providing new releases to improve the experience for your users, developers and administrators.
“I’ve purchased many platforms and applications but never been as satisfied as I was with OpenChannel. Never let us drop the ball. Super quick to deliver. Always responsive. Quality work. Amazing platform.”
– Sean Kester, VP, Product Strategy at SalesLoft
You Can Still Build An App Store, But Should You
At the end of the day, it’s still possible to build your own solution but we believe the most important question is “should you build it?”.
Every minute of an engineer or product manager’s time has an opportunity cost. It’s not about what’s simply important, but what is most important. What brings customers in the (metaphorical) door paying, and that’s core product.
When a core product has product-market fit, the user base continues to grow…and it’s that access to users through your app store platform that will attract developers to build apps and add-ons on top of your product for years to come.