Who builds the first audio search engine?7 min well spent
A couple of days ago, I opened my Apple podcasts app and saw this.
Apple beta tests generic audio search. We’re not talking about finding podcast names or hosters, but also guests and topics that are covered. Exact match search is not a big challenge but creating a search engine across many taxonomies isn’t easy. Google can sing a song or two about that. Now, Apple seems to expand its capabilities of adding generic search to audio content.
The opportunity is huge. I wrote about it several times. Most recently, in “podcasts in search – a new hope”:
Smartphones, AirPods, and smart speakers gave podcasts a second chance. There are over 700,000 free podcasts now. Listeners spend +6h per week on podcasts. That massive push brings the need for search with it. How are you going to find what you’re curious about in that content jungle?
The market is growing rapidly (ad revenue is supposed to top $1b in 2021) but there is not real “search engine” for audio content, yet. You can search for show names and hosts but not topics, guests, and other taxonomies.
Several challenges arise when creating a true audio search engine. You need to build an index based on audio transcripts. You need to rank content in search. You need to get as many podcasts on your platform as possible. You need to customize the experience for users. Etc. Etc. There is a lot.
To me, there are only three companies in reach of making this happen right now: Apple, Google, and Spotify. Who has the best chances to make it happen? Can Google defend the throne? Will it pass the torch to its neighbor from Mountain View or the foreigner from Stockholm?
In “Podcasts in search – a new hope”, I mentioned:
Companies like Luminary, Brew or Podimo want to tackle the problem and now have to face Google as competitor. Andreessen Horowitz invested $15M in Substack. Google entering the space is no good news for some of those startups, but it is for marketers and podcast creators.
Many companies are going for the holy grail but not one comes close to the most valuable consumer brand in the world: Apple had > 525K active podcast shows (source) in 2018.
According to Anchor, this is how the podcast market share cookie crumbles right now:
*An a16z article describes Apple’s market share as closer to 63%
While podcasts.apple.com still has an outdated look…
Google recently updated podcasts.google.com.
But here’s the kicker: podcasts.google.com has only ~200 links to other podcasts or categories. Podcasts.apple.com has ~200 links to categories, which then again have another ~500 links to shows. In other words, Apple optimizes its internal linking structure much better than Google.
That’s why Apple has the upper hand in SEO: According to AHREFs, podcasts.apple.com draws ~2.3M visitors and has ~130K top 10 rankings. Google has ~65K visitors and ~700 top 10 rankings.
Spotify gets ~55M organic visitors – even after losing 33% traffic after Google’s core update – and is the strongest player in this category.
Organic Traffic (according to AHREFS)
Apple: 2.2M (podcasts.apple.com)
Spotify: 34.3M (open.spotify.com)
Google: 65K (podcasts.google.com)
Google has the advantage of already being a search engine and software company, arguably the most successful one in history, and has the technology to index video content. This means Google could make podcast episodes even more prominent in Search to drive the product forward.
In fact, two days into writing this post, I received an interesting email from Google Play Music stating that the service will continue as “Google Podcasts Manager” because Google’s music offering will be transferred to Youtube Music.
Apple, on the other hand, is the most valuable consumer brand in history and the largest podcast hoster. iTunes is already a huge platform and could transfer lots of listeners to podcasts.
Spotify did the same thing: they bought Gimlet media and a couple of other businesses to kickstart their podcasting platform. The overwhelming SEO performance and their customer base are huge advantages they have over Spotify.
All three companies also have music offerings that provide huge carry-over effects. Let’s compare the Monthly Active Users to get an idea who’s leading here.
Youtube Music: 20M users (source)
Apple Music: 60M (source)
Spotify: 271M (source)
A huge advantage that Apple and Google have over Spotify is that they’re selling phones that come with a pre-installed podcast app.
Looking at Apple’s history, building a search engine wouldn’t fit in. Granted, Apple has been adding more service components to its offering as hardware growth has slowed down. In Apple’s defense, they hired John Ginnandrea in 2018 from Google to lead Machine Learning and AI Strategy.
Google seems like the candidate that intuitively has the best chances to take the lead here but seems surprisingly passive when it comes to audio content. The company seems to be more interested in video, which is fair because its a much bigger market than audio. I did write about the growing revenue share of Youtube and the social video war – both of which Google is very engaged in. That could distract from podcasts, though.
Spotify’s problems could also stem from the tech side. The company has created a fast and reliant platform but building an audio search engine is a different beast. I haven’t seen any signal of progress in this area, so it’s hard to judge.
The field seems to be pretty even here. Spotify wins in SEO/traffic and MAU but doesn’t seem to be interested in building an audio search engine. Google has the best technical resources but loses in market share, SEO/traffic, and MAU. Apple wins in market share and covers the middle of the field for the two areas.
My personal opinion is that Apple has the biggest drive and chance to actually win here. An audio search engine would create an attractive moat and user acquisition loop that fits the business, could be created with the new focus on services and brainpower, and the user base is there.
However, in the end, everything is open. When Google became successful, nobody thought that another search engine could kick Altavista, Yahoo, or AskJeeves off the throne. And look what happened.
One thing is clear: whoever figures out the advertising problem in podcasting will make a ton of money. Podcast ads currently monetize at only 0.01$ per listener on average (source). But if a company were to figure out a way to allow programmatic ads in podcasts, similar to how you bid on Facebook or Google ads, this number could drastically grow and create a new ad market out of nowhere.