Google has been the dominant search engine of choice for a long time. It’s the market leader. It’s the default for so many people on so many platforms. Everything Google has done over the last 20 years has been consolidating their dominance and making them the defacto go-to for their users for as much as they can. It’s all about aggregating data and attention to understand users and bring things together. Google maps, for example, providing you directions is not directly monetizable, however, it boosts the data they have on you to help make search more accurate.
In order to achieve what they have, they have been using AI (mostly Machine Learning) behind the scenes to power and improve everything they do. Looking at “classic” search, that is only possible to the level Google do it with massive Machine Learning and Natural Language Processing technology behind the scenes.
First, they crawl every bit of the internet. Then they put that text through Natural Language processing to see what that text is really talking about. Then they look at your search. You searched for ‘mattress’ - do you want to buy one or learn about them? So they then have to monitor search behaviour. What are people seeing in their results based on relevance? What do those people do to clarify the intent of their search. That’s what all that “People Also Ask” is about - are you looking for the “how is a mattress made” or “buy a mattress” really? They’ll also use their machine learning models trained on your own personal data and behaviour to work out their best view on what you’re after.
For example, if I search for “Elastic” I will get search results including the search technology platform. Other people who aren’t always searching and browsing in that space probably won’t see that on page 1.
Google’s recent webinar was announcing a lot of “exciting” new changes that Google are making enabled by their investment in AI to improve their services. But the truth was most of those changes were increments on what Google already do in that space.
The problem is with that massive, long term investment, Google as the incumbent unassailable market leader can’t release a beta or prototype thing into their mainstream channels without significant impact on them. As shown by an error in an early screenshot of a Bard chat dropping billions off their value.
The problem with that incumbent position is that despite the fact Search Engine Results Pages on Google are so heavily gamed by Search Engine Optimisation Executives in so many businesses they are increasingly full of rubbish. Plus, as Google would have to go really hard to impact their market share in Search, the amount and size of the paid advertising features are taking more and more space. It’s getting worse searching on Google. But that’s not denting their share of market as no other player has their reputation, trust and engagement in users heads.
Microsoft and OpenAI on the other hand have little to lose. Bing is the default search engine of the default browser of a Windows install. Only windows users with a corporate lock down in favour of Edge or not enough technical knowledge or interest stay with it. Everyone else still moves their search engine to Google.
Microsoft might be doing a lot of what Google is doing, behind the scenes, with Bing. But I don’t think it’s dragged in things like Navigation, Office etc into that space. It’s search seems more of a silo, and it’s just search.
They have room to make big risky plays. Their users are users by default that can’t escape. They have a tiny share of users by choice. They can do anything and if it doesn’t work it won’t make things worse. Many businesses that invest heavily in their Google Search results monitoring and their Google Paid Search spend barely bother to look at what’s going on in Bing. So they’re not going to lose any of that traffic.
They can do anything, but they can’t compete with Google on Google’s core search features and mechanisms. Google are so far ahead they can’t catch up.
Enter ChatGPT. It’s a totally different use of AI tuned to give a better more interactive engagement of Search. Google talked big on their call about all they were doing to help people search better, but they’re too locked in to their 10 blue links + ads model. They don’t want to lose that focus on businesses investing in their Google Clicks.
Microsoft want to win traffic and work out the rest later.
ChatGPT is someone really thinking about how users discover information, it’s looking at the clumsiness of searching, refining, clicking into websites, bouncing back to the results on Google and trying something else. Reading the same awful content dumbed down and tuned for search that doesn’t really answer the question as a result.
ChatGPT reads all of the internet content on a subject, parses it for meaning, understands it and summarises it back. And you can then ask it more about things. That’s how people learn, understand and research. It gives them a pseudo-expert source to discuss.
Yes, it can get things wrong, because it can misunderstand the internet, and the internet is often full of the wrong answer. But you could have got the wrong answer off a web-site.
Microsoft’s demos of how this will work in Bing, combined with the integration in Edge is outstanding. It really shows them thinking about how this can help users. Die-hard techies who’ve always been on Google have signed up for the New Bing waiting list, installed Edge and switched to it in anticipation. It’s a big deal. Used intelligently, all this stuff can make people’s work a lot easier.
Now, what’s really interesting is Google’s reaction to this.
It’s nothing short of panic. They whipped up that demo they did. Which was a disaster. Missing phone aside, people who knew and cared about this stuff saw it as the minor advancements it was in the round. It’s stuff they’ve been doing in public for ages. Combining their existing AR Google Maps directions with quicker integration into Search isn’t going to change anything for anyone.
They also announced and demo’d their Bard solution, which looks like a rapid acceleration of a pivot of their plans to do exactly the same as Microsoft/ChatGPT. Microsoft have had longer with their ideas in private. ChatGPT was a startup with no direct mass-understood link to Microsoft. They released a cool thing that techies, content creators and digital-first people rushed to play with. Microsoft could let that ride for a little bit. Establish it wasn’t a massive problem. Establish that it did work.
Then Microsoft could come in, PR their relationship, gear up, and announce their plans they had in the background.
It’s a shockingly immature response from Google to rush out with a “look we’re doing the same!” and “we’re more AI than them so ours will be better”. It looks like a panic’d reaction focussed on a short term spike in interest in Microsoft/Bing/ChatGPT/OpenAI.
A more mature, infinitely minded mindset approach would be what I’d have expected to see from Google. Sure, short term, they might loose a bit of mind-share to Bing and Edge (which is just Chrome in a Microsoft mask anyway). But longer term, their heavier AI guns would have inexorably blown Microsoft back out of the water.
That’s what’s most fascinating. Why have Google reacted like this? It’s more in-line with the way Elon Musk is changing Twitter than the deeper, longer game Google have been playing for 20 years.
I don’t get it.