Half of the MightySignal team is based in the United States. Those who live abroad are in Europe and South America. As the coronavirus raged through the world over the last year, I've been able to witness second-hand the ebbs and flows of infection rates and lockdowns and now, fortunately, inoculations. The world has made great progress, but the benefits of that progress have been concentrated rather than dispersed. As life actually starts to return to normal in the United States, my colleagues in Europe are fighting back infections with more severe lockdowns. Things couldn't be more different.
It's with that backdrop that I think about what we've done at MightySignal over the last several months. I recognize it's been a while since I released a summary like this -- I'm sorry about that, but I'm here to share one now. As I reviewed our journey so far this year, I couldn't help but note that while we've made fast progress in some areas, in others, it's been agonizingly slow. And yet we move onward! Like the global fight against the coronavirus, we're taking it day by day, app by app, code push by code push.
SDK scanning on Apple Silicon
We dramatically changed the way we do iOS scanning and the results have been phenomenal. We first described it here in our post about SDK scanning on the Apple M1 and our new Mac Compatibility value on our App Details feed. Now when an app is MacOS compatible, we use a machine with Apple Silicon to download and scan the app. It's far faster, more efficient, and less prone to error than using an iPhone to do the same calculation. When an app requires iOS, we send the scan request to our iPhone lab, which also has had numerous upgrades.
The net result is far, far fewer Live Scan failures, which have historically been our top complaint. We're glad to be the only SDK company to offer Live Scans, but it opens us up to critique when it doesn't work. And sometimes it just doesn't work! Fortunately, the failure rate on MacOS is much lower - and a majority of apps are now compatible on both MacOS and iOS, so successful scan rates are steadily improving.
Defining SDKs using Swift and SwiftUI
Coming up later this year is a rewrite of our SDK classification system. We recognize that there are differences between Swift and Objective-C classes and symbols. Our old system only recognized Objective-C classes, and as iOS developers release more SDKs using pure Swift and SwiftUI, our fidelity will decrease. We can't allow this! So our new system identifies all methods, classes, structs, types, enum cases, static variables, thunks, properties, protocol conformances, and more - for a total of 489 data points - that we can use to detect the presence of a specific SDK.
Backfilling scans of all iOS apps
What this ultimately means for MightySignal customers is we can scan a LOT more frequently. In fact, we've initiated a process over the next few weeks that will scan every single one of our iOS apps. There should be no old scans in our system when this is complete, and then going forward we'll be able to keep up with every new app release essentially in real-time. This will mean you won't need to Live Scan because we'll be caught up on all releases. I believe this goal is achievable in the next couple of months.
Improvements to our ad intelligence
Huge strides have been made elsewhere in our codebase as well. We fixed up a lot of old code in our ad intelligence system that had gone stale and was not pulling as many ads as it could have. You'll now see 2-3X as many ads in MightySignal Parade and our ad intel features in the MightySignal platform. We expect to add a new ad network to our system in the next couple of months, too.
More and better publisher contacts
Finally, another area for MightySignal to improve is our publisher contacts. Many customers rely on us to not only share what publishers have SDKs installed, but also who to contact at those publishers. Our contact discovery widget hosts millions of contacts, but the emails are stale and it's not kept up to date when people change companies. I am working on a solution for this problem aim to have it released by the end of the second quarter. It's a solvable problem and we have the tools and connections to do it better than any other SDK intel platform on the market.
So that's the summary of where we're at! As always, we'd love to hear from you. Contact us any time.