The Dream of Spotlight Search Is Still Unrealized (and here’s why)
It all seems so simple...
Type in a query, see relevant results. That was search, be it mobile or desktop. At this point, it’s taken for granted on desktop. It’s just there, and it works. (Thanks, Google!)
But for so long, pages inside iOS apps remained obfuscated behind a veil of functional passivity. Apps had tons content pages, but there were few ways for developers to put those undiscovered nuggets in front of their existing users’ eyes. As a result, apps were only as good as what the users’ already knew they could do.
Then came the big announcement. On a cool July afternoon in San Francisco, Craig Federighi proudly stood on stage and introduced to iOS users to their new best friend... Core Spotlight Search would finally unlock our ability to find content on our phone at the OS level, alleviating the need to manually search each app, or use a jenky Jerry-rigged third party solution to find relevant content.
This exciting new development was not only going to improve our ability to find content, but also make our mobile experience more powerful. We’d be more engaged with our apps, we’d have the ability to maximize each app’s potential, and save time and effort to get to content. The red carpet was being rolled out.
Again, it all seemed so simple!
Next came the documentation.
Apple had done us all a favor. Not only was this going to be a massive functional improvement. It was also super easy to implement! Just associate some meta data with your pages, and decide what you want to pass to the device index at any given time. Great. Simple, clean, very “Apple.”
Apps began to add the relevant code to their framework, and get ready for the Fall launch.
And then it launched... iOS9 was here, and all the sudden, the badge app icon showed 30, 40, 50 apps needing update (for those of us who toggle auto-update off, that is). For those of us who make our living in the app engagement space, it’s always enjoyable to check out the update notes when a new OS version launches. The notes talk about what the app now supports, and updates they’ve made (except Medium app, which is more a Slack chat transcript)...
If you’ve been watching those release release notes from September on through today, you’ve inevitably noticed that they all share a certain quality in common...
Here, take a look at a few examples:
RIGHT! Gold star for you... Apps are leveraging Spotlight to allow their users to search for content pages they have previously found inside the app. It’s a quick path back to a page you’ve already seen.
“But Google lets me find anything I want on the web!”
And therein lies the fallacy in comparing Spotlight Search to Google’s algorithmic search platform. Google made their bones by helping users find the best content from across the whole internet. It brings to mind an old quote from Eric Schmidt:
“When you use Google, do you get more than one answer? Of course you do. Well, that’s a bug. We have more bugs per second in the world. We should be able to give you the right answer just once. We should know what you meant. You should look for information. We should get it exactly right and we should give it to you in your language and we should never be wrong.”
Now I suspect Mr. Schmidt was employing a bit of hyperbole, as Google makes their money by showing a number of advertisers in a given search. But even still, the lesson is no less relevant. By design, Google is the layer of search intelligence and decision-making.
Their philosophy dictates to developers: you bring the content, and we’ll will crawl, index, and surface it based on relevance” (and maybe some other stuff). From day-1, Google theorized that making results incredibly relevant would bring users, which would bring advertisers, which would bring revenue (hindsight note: file that under N for “Nailed It”).
Apple’s world is totally different. They don’t need to worry about serving the best results. They needn’t build their search products as a conduit to new users, and they could give a hoot about search advertisers.
Apple has myriad products that force us to open our wallets, irrelevant of the ability (or lack thereof) to search for content. What they need is a serviceable user solution, one that allows apps to get some (any?) content results in front of users. And that’s what they’ve built. There’s no heavy search algorithms or predictive analytics, and no obvious data mining or machine learning to improving results over time. Whereas Google only needed content, and built the surfacing and optimization technology themselves, Apple is farming that out to the app developer.
Says Apple: You provide the content and then tell us what links each user should find.
So, long story short, there are inherent limitations with Spotlight Search, as it exists today. Most notably, any given app can only pass so much data to the index at any moment in time. This alleviates the need for the device to do heavy local result optimization, and keeps the search lean at runtime. And because the app now will need to decide what info should be found by each user, every app now needs to build a strategy around what content should be indexed, found, and served.
So, naturally, app dev teams (which are notoriously hamstrung by unmanageable backlogs and limited resources) are defaulting to showing what checks both the “easy” and “logical” boxes...
And thus the, when you break through the heavy press and shallow accolades, you realize the truth about the first production release....
Spotlight Search = Bookmarking.
So how to make it better?
You can’t... (Debbie Downer)
OK, that’s not totally accurate. You can! But you’re going to be limited for now without some crazy internally product development initiative. What would make Spotlight Search better is a tool makes intelligent decisions (in real time) about what to send to the device index. It needs to understand what content is on a page, what the context of that entity is, how to get to it (which is thankfully becoming much easier as more apps integrate Universal Links), and have it organized in a simple and easily-accessible taxonomy.
In the web world, we know a congruous solutions as Search Engine Optimization, or SEO. The premise of SEO is to optimize the content and meta data of a web page to make it easier for Google to crawl, understand, and serve as search results. With Spotlight, the end goal is the same (serve the user the most relevant results), but the flow is unique, the signaling is totally different, and the technical necessity is wholly unrelated.
Traditional SEO requires an understanding of Google’s crawlers; what they look for, what data they value, and how to “massage” a site’s markup to make sure that those checklist items are hit. Google still does the actual crawling, indexing, and matching. Core Spotlight (as noted above) offers no layer of search intelligence or optimization, so the app framework must do the heavy lifting to pass to the device links that might be relevant for the user.
By comparison, imagine if Google’s documentation dictated to web developers that they should “tell us, each time a user goes to google.com, what content pages we should serve to that user.” We’d all still be using Yahoo! directories.
We need an optimization tool that can work within the Spotlight guidelines.
Well, not exactly. It’s more Spotlight Optimization than Search Engine Optimization. The idea is, apps need a tool that will ingest and understand the content of the pages inside of it, and then use the vast amount of signalling that exists on each device to return relevant results. More specifically, this tool will replace an app’s “Spotlight Search Strategy.” Apps won’t have to decide to pass previously visited pages to the Spotlight index. Instead, all the app will need to do is toggle on the ability to be found in Spotlight searches. The Spotlight Optimization tool will do the rest.
Every so often, the founding team at Deeplink looks back at the original Google AdWords Press Release. It never ceases to amaze us how beautiful the product is in it’s simplicity, and how the core offering hasn’t changed. Advertiser buys search term, and search term brings advertiser link.
In a world of disruption and innovation, AdWords still stands alone as one of the “I must do that” solutions for every company, regardless of size. And the beauty for the AdWords launch was that Google had a large channel for distribution already baked in, who were declaring their intents at any given moment.
Today, device-based dynamic search stands at a similar precipice (be it user-querying, or preemptive search and discovery models like our AppWords platform). Apple has us users raring to go and wanting more. Though their business model is different, the users are indeed primed. The ability to improve and smarten up the results is an immeasurable opportunity for whoever can crack it.