Google Instant Steals 25% of Your Organic Search Results | Hacker News Reader

Google Instant Steals 25% of Your Organic Search Results

blog.keltex.com

68 pointskeltex54 comments

There are two blog posts with real data from tens of thousands of search queries showing that in reality, the impact of Google Instant has been negligible:

- http://www.seomoz.org/blog/google-instant-fewer-changes-to-s...

- http://www.distilled.co.uk/blog/seo/impact-of-google-instant...

I wish this type of sensationalist supposition without data to back up the assumptions received less attention here at HN. I've always been impressed with the culture of "show me the stats" when it comes to subjects like financing, startups, technology, engineering, etc. Not sure why that dissipates when a marketing topic arises.

Well, the impact on search query length may be negligible but there are other ways in which Instant may be changing search behavior. I have a small slice of data I hope to post in the very near future.

Here's some preliminary data that shows that Google Instant may have changed the distribution of clicks by rank, with more emphasis being put on results above the fold.

http://bit.ly/dAzvX3

It's not about the query behavior, but about assessment behavior. As such, your mileage may vary based on your particular rank for terms.

Outside of this analysis, the keyword length could remain the same, but a site may receive less traffic if the auto complete algorithm has changed for critical terms. Again, this will vary site by site - and where one loses, another will gain.

Well, Stack Overflow gets about 90% of its traffic from organic search results, on the order of millions of daily page views. The traffic Google sends us did not perceptibly change when they switched to Instant. So anecdotally I can tell you that this is just demonstrably not true for one large-volume site.

Now that I think about it, the whole argument is bogus. For every SERP where your result loses traffic by virtue of falling below the fold, there's probably a SERP where your result gains traffic because the competitors fall below the fold.

Sorry, this hypothesis is neither logically correct nor is it backed up by any facts. Do some real experiments next time before you start making claims of having discovered scientific truths.

I don't think that SO's users are representative of google's as a whole. For example, there are a lot less advertisements on software development related searches. What I was pointing out is the high-volume, high revenue searches are being impacted. Plus I'm guessing that people who do those searches are much less savvy than your community. They are less likely to scroll, they have smaller screens and they are problem less discerning between paid and organic results.

Any evidence of that?

Which part? The fact that novice users have a hard time distinguishing between PPC and organic results? Anecdotal with the many novice users I have dealt with.

That there is less advertising for software related keywords vs. other types of keywords? Try out Google's adwords traffic estimator: https://adwords.google.com/select/TrafficEstimatorSandbox

Any evidence that "high-volume, high revenue searches are being impacted"?

In addition, the push down of the "fold" only exists as long as the user is doing the "instant" part. As soon as the user decides the search is what he wants and either hits <ENTER> or clicks on it with the mouse, the drop-down box disappears and the results that were "below" the "fold" go back to the original place.

The 25% statistic is also tenuous: if you have a bigger screen than the illustration in the blog, you will have more results shown (i.e. Google will "steal" less than 25% of the results). If you want to be more dramatic, do an instant search with a really small window and OMG!!! Google will "steal" 100% of your results.

Does Google drive a significant amount of organic traffic to Stack Overflow for short term searches?

I permanently disabled Instant after a few searches, so I have no opinion one way or another as to the effects of it on search. However, it seemed like it was designed to be useful for short queries. Stack Overflow, on the other hand, is extremely useful for long tail searches. To be honest, since I typically use SO to solve a specific problem, I don't know what short terms I'd use to land on a page of interest.

Instant's goal is to change user search behavior away from potentially long tail queries towards the head.

If they can short circuit your potentially 4-6 word specific search that has fewer and lower paying advertisers associated for a shorter, higher interest/CPC search, they stand to make a lot more money and please larger brand advertisers.

That they can wrap it up to users as "decreasing search time" and "better search results" is brilliant.

I don't understand the cynicism. Do you have any evidence for this whatsoever?

He's absolutely, screamingly right. They're an advertising company. Big bang changes to the website exist to sell more advertising. Everything else they do is window dressing to make people forget that they are an advertising company.

If all Google showed were ads, no one would come to the site. Furthermore, Google is already outrageously profitable. The only real threat is if people start searching somewhere else. Do you really think they would make changes that make that more likely? Google has strong competition in Bing, Yandex, and Baidu. Even if profit were the only relevant motive, short term profitability is not something Google needs to worry about.

It also seems that your argument applies to every ad supported site. Are you contending that the nytimes is in the ad business and not the news business? Is the news mere window dressing for their advertisements? If not, then how are they different from Google?

That's actually a better characterization of the newspaper industry than you probably mean for it to be. From the editors' perspective, of course it's all about the news (I'm sure this is true for Google's search engineers too). But from the publisher's perspective, the news basically is some very attractive window-dressing.

True, but it's Google's search engineers that made this change, not it's ad engineers.

Everything else they do is window dressing to make people forget that they are an advertising company

That's a bit disingenuous. In the same way, newspapers are also advertising companies. The journalism they do is 'window dressing'.

If they nudge people away from long tail queries, won't this mean that people will perceive the relevance of the search results to be worse?

I doubt Google wants to start driving people to other search engines by making it harder for people to find what they are looking for.

If they nudge people away from long tail queries, won't this mean that people will perceive the relevance of the search results to be worse?

Only if the long tail results were better. I've sometimes gotten more relevant search results from removing keywords, even when those keywords seemed to be on the page I was looking for anyway: other pages were better ranked for those keywords, but less useful to me. I can see an argument, therefore, that presenting the intermediate results on the way to the full query could sometimes mean better results, quite apart from the advantage of getting the result you want faster.

Some money-making things, besides higher CPC, are highly correlated with "decreased search time" and "better search results". Mainly, the fact that Instant (to some people) makes Google their preferred search engine. I think it's a rather large claim that Instant's main purpose is to change user's search behavior.

The core goal of Google, even as an advertising company, is to provide the best product: search results, and derivatively, your experience getting to those results. While what you are positing might be true, I suspect that core goal of Instant is to provide a substantial edge over Google's competitors.

If I were Google right now, my main concern would be: Other search engine's results are almost as good as ours, and for the large part, "good enough." We need to maintain our competitive edge by creating the best search experience possible: fast, novel [notice more doodles lately?], and with "innovations" features that generate buzz about our edge-cutting.

I'm working on a product that aims to replace Google as users' start pages. The goal has become: How can I make something that gets users to specific results faster than Google does? The answer: For a lot of things, there is no possible way. They've put so much into making their product fast and efficient, which is all you really care about when you're looking for something. I suspect Instant is just another edge they want to have over competitors in a such a valuable market.

In case you're wondering about what I'm making, the goal has actually become: How can the process of searching be commodified?

User's overwhelmingly click the top result anyway. [1] If you're not in the top 3, you're close to invisible.

I'm not sure this is the scandal the author makes it out to be.

[1] http://training.seobook.com/google-ranking-value

No more than having a small monitor steals organic search results.

No, it does steal more. Whatever your monitor size the instant suggestions should steal enough space to knock out an organic result.

It's a damn shame there isn't anything like, say, a scroll wheel or a scroll bar, to look at the results below the fold. That would really sort things out.

I bet small monitors DO have an effect, and small monitors + instant is probably worse.

But what that effect is exactly, that's harder to say. It's not quite as simple as the OP seems to think, IMHO.

There's been quite a lot of UX research on this, and the concept of "the fold" simply doesn't exist for webpages. People scroll and scan, even the least knowledgeable ones.

Not when they're in the process of typing.

When most people are in the process of typing, let's face it, they're looking at the keyboard and hardly at the screen at all, let alone, scrolling down.

You're exaggerating. Web users spend 80% of their time looking at information above the page fold. The part of what you're saying that's true is that people scroll at all. This is new. See http://www.useit.com/alertbox/scrolling-attention.html

And yet that page had at least 7 folds.

I think the goal of enabling visitors (and expectation of visitors) to utilize strategic navigational features and view key information within the top portion of the page is valid.

And yet, convention dictates web site navigation to be concentrated within specific page areas (top, bottom, sides) and most visitors can anticipate certain features and functionality to reside within one of those areas as best (popular) practices are copied and commoditized.

Search results are not the same as the loud flashy bleeps and blips on the latest MTV website. There's an expectation that a user will navigate through the page, on to multiple pages if the first page is either full of useless ads and/or SEO pollution or unsuitable expectations.

I personally believe 'the fold' of a web page is a bunch of crap. But that's just me

You're both right. Users aren't reluctant to scroll if they believe there is something they want to see. However, most pages are designed so that all of the relevant stuff is above the fold. Most of the time, you just don't need to scroll down, because the page is set up to have the most relevant things above the fold, and the least relevant things below.

We're both right, except for the part where he said that the concept of "the fold" simply doesn't exist for web pages.

Regardless of whether they're correct about it lowering traffic or not, I strongly dislike the use of the word "steal" here. To me that signifies a strange sense of entitlement; the search results are not "yours" and you don't have a God-given right to them, so Google cannot steal them from you. You might as well claim that "I'm Feeling Lucky" steals your search results because users never see any mention of your site if you're in second place, or that showing 10 results per page is unfair to those in 11th place.

I wonder a little if this reflects a general debasement of the word "steal". Some years ago it clearly referred to taking physical property, but maybe the smear campaigns from the music / movie industry have successfully extended the scope of the word to include things like this.

Interesting observation. It's very subtle, but potentially important.

As for myself, I don't think there's a big difference. I use Google Instant just to help me form my queries, because my query changes as I see the intermediate results. The paid results and the first organic results help me see if my query really is going to give me the most relevant results, otherwise I try modifying the query a little bit. When I'm done with that phase I click search, which removes the suggestion box, finalizes the results, and then I scroll down and open whatever results appear promising.

Do we know if the suggested results are based on search frequency or revenue maximization? -- i.e. if I type in "plan my" will it suggest "travel" above "vacation" if the PPC/CTR on the ads for "plan my travel" generate more $s than "plan my vacation"?

That seems like a much more clever optimization than stealing screen real estate. The other optimization to this is if they can guide users to a smaller list of common search phrases they can drive up the prices for them -- a fragmented list of search phrases is very bad for Google.

Interesting, so...

ORDER BY (CPC * CTR);

versus

ORDER BY (CTR);

yeah, i know they're not using SQL

CPC * CTR = RPM, and that's what AdSense optimizes for. In effect it's brilliant in it's simplicity as it combines both advertisers wishes (CPC) and consumer's wishes (CTR).

once you hit return, the suggestion box disappears, and you are back to normal mode, do people not hit return?

Google has two goals with Instant... first, they want to improve the user experience by offering inline suggestions - makes sense - I like that part.

What they're not advertising is that Google is dealing with the same decreased CTR and advertising revenue that the rest of the Internet is. In an effort to seem as user-centric as possible, they've introduced a user friendly feature with a refinement that increases the CTR on their ads.

This isn't a shocking move, it's a sign of the web to come. What's really disappointing is that Google isn't devoting more time to coming up with different ad units and testing more creative ways to encourage users to click. The online ad industry is stagnant - advertisers and publishers are scared to make changes. Google isn't any different.

My startup is building better displays ads for the web: http://nowspots.com

I agree Its weird that the basic deliverable of an ad unit, a click, hasn't changed in 15 years.

I've always found it funny that advertisers swear by CTR as much as they do. It's almost as if they completely forgot the last few decades of television advertisement. For television, no one gives a damn if a viewer goes out to buy the product right that instant. It's all about repetition and building psychological reinforcements so that once the viewer actually needs something along the lines of what that product offers, they subconsciously lean towards that specific product as opposed to its competitors.

Granted, it's a lot easier for someone to just click some text from their computer than to get up in the middle of a show and drive to a store, but the same essential theory still applies. Which is why I don't understand how CTR is valued so highly over impressions.

The 'Old Spice' campaign is an example of what you describe.

I think you're speaking in terms of a specific market. A local florist ppc'ing does not have the luxury of affording to brand in this economy.

inherently a click affords more branding anyway as you are much better able to control the presentation of your brand on a page you have complete control over.

>So instead of giving you “better search results”, you’re getting less search results.

Are certainly not mutually exclusive. And when searching for a specific thing/site (rather than going for broad exposure on a given topic) less results, preferably 1 (the correct one) is best.

I have to say that I am NOT a fan of Google's instant search. I hope that Google will one day offer a URI that has instant turned off by default, such as http://google.com/instantoff

Fast company article about Google instant and ad impressions -

http://www.fastcompany.com/1687976/google-instant-impression...

I'd like to see a poll of users to find out what percent actually uses instant as it is meant to be used. I know I still do searches how I always have, by hitting enter. I imagine most non-technical users are still hitting enter.

What percentage of people actually visit google to search?

It's much less "instant" if you add in the time to get to the google page.

I user the address bar in chrome and my browser search bar in firefox.

You're assuming that each place on the first page gets equal amounts of traffic. The 4th result does not get the same amount of traffic as the 1st result, it gets much much less.

I can say that both on my personal sites, and on my Fortune 100 client websites, I have seen increases in search traffic from Google after instant was lost, not decreases.

today is the day of science on HN - this 25% of Organic Free-Range Antibiotic-Free Search Results were definitely topped by masturbating squirrels. I bet squirrels had much more fun in their story than Google in its.

a lot of these "improvements" are made for making people pay. i.e. like that local business results

I can't remember the last time I used a screen whose native resolution was 1024x768.

Reply on news.ycombinator.com