Apple’s paid search experiment shows there’s still no PageRank for apps


Hey, over here! Paid ads need to be relevant. Photo by Michael Rehfeldt on Flickr.

So Apple is doing it: introducing paid search ads to the App Store. People will go to the App Store, search for something, and if a developer has bought an app ad (it can only be an app; no links to content outside the store) and it’s deemed relevant by Apple’s algorithm, then one will appear at the top of the results, backgrounded in blue and making it clear it’s not part of the organic listing. Apple has put up more of the detail.

Of course this has riled developers, for reasons I’ll explain. But a few things first.

There will only be one ad because, Phil Schiller told me, people are going to be searching on mobile, and “we don’t want to push organic search results too far down the list.”

Schiller’s rationale for introducing paid search goes like this. People who want to get apps find them through searching: there are hundreds of millions of searches every week on the App Store. Two-thirds of downloads come via searches. If you want to advertise your app to people who are looking for your app, or something like it, where would you want to advertise?

Logic would suggest you’d do it right there, around search. But until Monday (when it starts in the US, in beta) that avenue hasn’t been available. So developers have resorted to all sorts of tactics – social media, plying reviewers with downloads in the hope of good reviews, paying reviewers for good reviews (on some of the scuzzier sites), buying ads that redirected to the App Store (which I think indirectly drove Apple’s introduction of content blockers), trying anything.

So Apple says: hey, stop spending your marketing money where you can’t be sure anyone will see your efforts. Instead, do it on the App Store, where you know they’re searching!

The mechanics are pretty much identical to Google’s AdWords (the mechanism that puts up ads against searches on Google). It’s an auction system, where the winner pays only what the second-highest bidder offered (so you bid $4, I bid $5, I win but pay $4), and pay-per-click – you only pay if someone does click. No minimum bid, no exclusivity.

“We look at it as giving every developer the chance to drive downloads through marketing,” Schiller said.

Meritocracy has been delayed

This, then, is Apple’s answer to developers’ and users’ repeated complaints that “search in the App Store is broken”. The basis of the complaint is that when you search for apps, you get too many junk results for apps that aren’t relevant, or are outdated/un-updated, or which are straight-up ripoffs.

In other words, there’s still no PageRank for app search. But that’s what people really, really want. Developers and users want a meritocracy; by going for paid ads, Apple is instead giving them an oligarchy.

Ahead of the call with Schiller, I contacted various developers, and some users on the Above Avalon Slack channel (you have to subscribe; totally worth it in my view). I didn’t say that Apple had any changes coming; instead I just asked what three things they’d like to see improved about the App Store ahead of WWDC.

Top of everyone’s list? “Better search”. But what do we mean by “better”?

When I pressed Dave Verwer (who runs the excellent iOS Dev Weekly list) on this, he admitted that

“search is hard. However Apple has a huge amount of data not only on the apps that we buy, but on those that we use, where we keep them on our device home screen. I’d love to see Apple personalise search results in order to provide customers with more relevant results.”

James Thomson (of PCalc and DragThing fame) was also in favour of “better search and discoverability”. But this is a motherhood and apple pie response. How do you do it?

“I’d like to see old apps that haven’t been updated in years gradually retired from the store. I don’t want to search for apps and find ones that won’t even run properly on the latest devices,” Thomson said. “I would (unscientifically) guess that over half the apps on the store are ancient and broken and if you cleared them out of the search results, that would improve matters enormously. I think paid search on keywords is a terrible idea for indie developers and will only benefit the big companies with deep pockets, rather than the users. It will make the playing field even less level. Search should return the best and most relevant results, not the results that have the biggest marketing budget.”

That last is the strongest point. Schiller told me that nobody will be allowed to buy out a keyword; and you can be sure that Apple will have learned from the experiences of Google, where rows over ads bought against trademarks have been many and vicious.

Even so, I wonder if Apple is quite prepared for it. I think policing ads for scam apps which put in fake metadata is going to be a giant effort in its own right.

What about users? David, a user on Above Avalon, put it like this:

“Discovery is the big thing I’d like. It’s just like Spotify – they have all the music, but I still just use my playlists I built some 6-7 years ago. Then they launched Discover Weekly – and finally it was a format where I could truly discover new music again. I feel like being sixteen again (I’m 32), finding bands and even entire genres.

“So if Apple managed to actually get me to download new apps that are not just “my bank released a new app for managing my index funds” or “this city council has their own parking meter app” – I think they and app developers would benefit. I rarely these days truly discover new things in the store. I doubt it is because new things aren’t released. They don’t even have to be new. Just new to me.”

Or ask Daniel Jalkut, another developer:

“I think for discovery, there is a great potential in tapping social trust networks. I know Apple is famous for “not getting social” but imagine if there were an incentive to both review and rate apps because people trusted your point of view, and there was some payoff in the form of fame or fortune? I think Amazon gets a bit of this in the fact you can rate reviewers and they get some kind of “top reviewer” status after a while.

“Similarly, what if I could click a little ‘trust’ icon next to Charles Arthur’s review byline, and from then out whenever I searched … for anything … apps you had rated well floated up? I would click the “trust” icon for friends whose tastes I share, prominent bloggers who I’ve seen thoughtfully review apps, and random strangers whose reviews and rating keyed into my same tastes. By having some kind of opt-in trust system, you would reduce the risks of gaming, because nobody could game their way into your trust network except by your approval.”

The rudiments of searching

As Bloomberg had already discovered that Apple was thinking about paid search, developers have had time to ponder what might happen. Marco Arment was unforgiving back in April:

Such a system would exacerbate much of the App Store’s dysfunction, disincentivizing improvements to organic search and editorial features while raising the cost of acquiring new customers above what many indie developers and business models can sustain.

But then he seemed to relent:

Assuming the system would be auction-based by keyword like Google AdWords, for less-contested keywords, marketing apps could become much easier. Buying a few good phrases could inexpensively put your app at the top of the list to help you get off the ground and start to seed organic growth.

More significantly, we could buy increased exposure to the most likely customers to buy our apps. More paid-up-front apps could become viable, and prices could rise.

What’s almost certain to happen is that the money that used to flow into social media campaigns and ads on various other media will instead flow to buying ads on the App Store. Apple thus will capture more of developers’ marketing budget. And (per the point above about the 65% of installs) it can argue that that’s as it should be. I would guess that Facebook and Google are likely to be the two who won’t be significantly hit. (Google might lose a little.) Ben Thompson says the same – Facebook will be fine. Update: Thompson backs that up with an excellent point: Google has offered paid app ads in Google Play for a year already, and that’s had no appreciable effect on Facebook’s app install revenues, even though Android has the larger number of downloads overall.

Just to reinforce that, one of the developers I spoke to said that they use Facebook to target people who they know will be interested in their games; the way it can deliver the ads to the right demographic works for them.

If Apple selling paid search ads skims off those scammy ads which take over mobile pages and dump you in the App Store – looking at you, deadline.com – then some publishers will lose out, but other and better ads can replace them.

What it isn’t: “better” search

This isn’t the PageRank for apps that people had been hoping for. But the problem is that despite so many people thinking and talking about the need for “PageRank for apps”, we still don’t seem to know what it looks like.

Is it downloads times activity? One developer I spoke to recalled a time when their app was downloaded millions of times in a single weekend; when they looked on the Monday, their app ranked in the late teens. “I may be biased, but I’d think we should be No.1, because we know people were using it,” they said. Sure, that sounds reasonable. Downloads? Rate of increase of downloads over a minimum? Activity per download? Apple can get all those numbers, as indeed do a number of the meta-services like AppAnnie.

I asked about this. Schiller replied that simply biasing search towards download numbers times activity, and not having ads, would mean that the big established players would remain. (Think of Instagram and Facebook.) There wouldn’t be a way for small apps to break through. There’s some truth in that, certainly. Perhaps there just isn’t a PageRank for apps. (Sameer Singh at App Annie reckons that Google Now on Tap, in Android 6.0, is going to turn into PageRank for apps, but thinks it’s “a few years out”. We’ll have to wait and see.)

The other stuff, with subscriptions, is potentially going to help a lot more companies achieve long-term business success: halving the take by Apple in the second year of a subscription is helpful. There’s no more data sharing, but at least there’s more money. Plenty will be happy with that, at least.

Another point to consider: how will this be done? I wondered if this will take an iOS update to achieve, since the App Store isn’t decoupled from iOS in the way that Google Play is from Android. Schiller demurred on this. There’s more to come at WWDC. The really amazing thing would be if Apple is going to decouple bits of iOS from the system apps, as Google does. That would be remarkable. But now I’m really speculating.

Finally, why announce this now? Schiller said it’s because there’s “so much” to come. Well, sure, but there always is; content blocking, which arguably is huge, wasn’t in the keynote speech (except as a line in a word cloud on one slide), and people only slowly came to realise how important that was during the week. Apple could, for example, have preannounced that. But didn’t.

No, I think that Apple saw how concerned people were about paid search ads when the Bloomberg story came out, and decided that rather than having the entire discussion post-keynote be about that, they would instead announce it formally, along with improved subscriptions and the already-happening “faster review”. Let the storm clear, and then move on.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s