Do Long Pages Work Better Overall?

In a recent discussion CRO was brought up (Conversion Rate Optimization) which for the most part is not a direct consideration of SEO since all the efforts you make don't directly impact on ranking changes (obviously you have the traffic to begin or you have nothing for CRO to do)

But that discussion suck me into the blog http://www.conversion-rate-experts.com/blog/ and there I found the trials of SEOMoz when the relaunched their tool campaign.

Reference: http://www.conversion-rate-experts.com/seomoz-case-study/

While you can't deny the obvious improvement in general I'd want to look more closely at the steps and appreciate whether there was any actual causation or did it just come along for the ride.

The blog goes through 8 main changes:

1. We created a web page long enough to tell the story
2. We infused the headline with curiosity rather than overt “buy me” language
3. We explained precisely what customers would get at each level
4. We showcased things customers cared about but SEOmoz had taken for granted
5. We augmented the message with video
6. We lowered the “risk bar”, to make it easier for customers to try the service
7. We asked non-subscribers to tell us why they didn’t even want a $1 membership
8. We created a quick-start guide so customers could get quick wins

The first one BLOWS ME AWAY!



The highlighted point I noted is my first concern... IF IT IS "What counts is not how long your page is, but how engaging it is." then how can you turn around and say longer is better?

In contrast. The 7 Deadly Sins of Landing Page Design by Tim Ash http://www.youtube.com/watch?v=erdEZvOq6wo suggest with #4 TOO MUCH TEXT IS A KILLER

What if the long page killed buyers and the headline & video increased buyers... but instead of making $10 million they only made 1 million.

Tim Ash - "study after study" show long pages are conversion killers ... and Conversion Rate Experts don't actually demonstrate long pages work better... they suggest the more obvious "engaging pages work better".

So wouldn't a short ENGAGING page work better than a long ENGAGING page... since most people on the web have a very short attention span.

But I really don't know...

any thoughts?

I think it is quite possible to have both..... so long as the long page can "keep" the user engaged. It is easier to keep someone engaged on a short page, but if you can do so on a long page, then I can see how the long page would be more effective......

but keeping someone engaged that long would be a major challenge, and I'm a betting that it would not be possible on most products. It seems to reason that a long page would suit products with a higher price tag, while something that is considered to be inexpensive would only need a short page.

Something that someone is spending what they deem a small fortune on would most likely require a longer page. If you keep their attention through half the page, but then lose them before they have enough info to make an informed decision, then it is a waste. In that case, you would need to either make the rest of the page more engaging, or find out how to get your points across in fewer words.

On the reverse, the cheaper product that wouldn't require as much information and gets by using a shorter page (yet still engaging), could really go either way (so long as the page remains engaging). But usually with a "cheap" product, there isn't as much to say.... and therefore the "engaging" part would be rather difficult to achieve.

So..... I like to stick to the "bathing suit" way of doing things..... make sure it's long enough to cover the subject, but short enough to be interesting. This will vary depending on what you're talking about.

Ya - I'm amazed that they talked alot about split running - but they didn't to that... there was the old way... and then the new way... and "ah this is good enough".. "see it worked"!

The pages on my site are educational, I don't sell anything. I'm not trying to hold anyone's interest, I'm simply providing information—if they want to know about it they read it. Most of the pages are 1500-2000 words, many are as longs as 3000-4500 words. None of it is funny, all of it is very "matter of fact." All pages are structured the same way and have multiple subheads, but no table of contents.

At some point I thought about how much is enough to tell them. At what point will they quit reading? Will they give up because the page is too long? So I decided to add a poll to the bottom of the most popular pages and see how many answered the poll assuming if they read to the bottom of the page they would participate. The poll was binary—yes, I found what I was looking for or no, I didn't find what I was looking for.

I don't know what the statistics are for people answering polls, but in my opinion the percentage of people answering my polls was low. So on a couple of pages after the first paragraph I added in green - "please don't forget to answer our poll at the bottom of the page." On those pages, the percentage of people answering the poll jumped up significantly—although the average time on the page did not change. In 100% of the polls—both the pages where visitors were reminded to answer the poll and the pages where the poll was not even mentioned—the visitor said they found what they were looking for.

So I decided to keep doing what I'm doing and assume that the visitor reads/scans until he finds what he is looking for—probably similar to what I do. I read introductory paragraphs, and then skip down through the headings until I see the heading that probably contains the content I'm looking for. I almost always scroll all the way to the bottom, regardless of length. I rarely participate in polls, or social buttons—unless the article just blows me away. I always participate in feedback because I sympathize with the website owner.

I hate short pages of a 2-3 paragraphs where the content is buried in the middle of tons of advertising and I can't tell where the content is headed or how many pages I'll have to click through to find what I want. Those I usually abandon and move down the SERPs.

Interesting how just simply noting the poll at the bottom increased the number of answers you received. I wonder if you could increase sales by saying "don't forget to purchase ....... at the bottom of the page"

I know the Conversion Rate Experts personally, We've been a client of theirs for about a year.

They DID split test the SEOMoz page they designed against the original short page, so the gains we incremental and not coincidental.

They know their stuff, we've seen sales more than double on the same traffic levels since working with them and the graph is still climbing month on month.

The key thing is that it's all about split testing (so you can compare like with like), and making radical changes, and seeing what does and doesn't work.. and then repeating, so no page or tool is ever "finished", it's constant improvement.

It's a lot lot lot easier to double conversion than it is to double traffic, and even more beautifully it's not against anyones terms of service!

Whereas before our growth focus was primarily on SEO, we've now switched to CRO.

And the beauty is that CRO is actually a process that tends to improve your website for the customer (a sucessful sale tends to mean they got what they want) which tends to increase incremental benefits like referals and even SEO positive factors such as links.

Few of the products that are at the top of Clickbank for long time have short pages. But those pages do have a sequence... The call for action will be at the end of the sequence. In a way this is also a long page. But they don't want to risk keeping all the content in one page as people may not want to read lengthy content unless its very important to them. They split the content into sequences. (like the tutorials at Seochat).

IMHO, a long page with just text may not work. A page long or short if it has elements like images, videos or the writing style which will make the visitor to temp to read the next paragraph will work!

"Split run" is completed on the back of PPC campaigns... having 2 or 10 versions of a website "non-indexed" is possble... obviously having 2 websites indexed organically and both ranked is impossible to do... and organic short term runs wouldn't be accurate.

While I do see the fit... my original point wasn't about CRO validity but the choice... paging vs. scrolling...

Study after study show paging (short "ON POINT" pages are the way to go for sales)... allowing those that need more to get more but not making those that need less to search endlessly through the data...

I'll agree going from unengaging to engaging improves things considerably... but look at Google search results... an "ENGAGED SEARCHER" never using depth 100 (which is about the length of the SEOMoz page.

SURL 2002 sample study on Google itself showed that 50 depth was most preferred and short was always better than longer. (that was a small sampling)

I rarely go out of 10 depth myself... course I could be an oddball.

You can run multiple versions of a single site at the same time:

www.google.com/websiteoptimizer

How it works is that it allows you to create multiple versions of pages on your site. Google will only rank and index you on the version you say is the original, but allows you to create multiple variants that it won't crawl or index (it's basically a form of cloaking, but done in a way they deem legitimate).

There's a small nub of Javascript you drop on your page, and then when you activate the tests Google Optimiser starts splitting your site traffic across the variants you've setup and tracks them for you.

This isn't just a PPC thing, it's ALL traffic.

So you effectively get your site to run multiple versions at the same time (meaning you can discount seasonal variations in user behaviour and traffic sources) and track performance of each variation against the goals you set.

Google optimiser isn't the only piece of software that lets you do this, but it's a good option.

This is all Conversion Rate Optimisation for traffic once gained, it's not about trying to optimise traffic for conversion.

The best thing about running tests like this is that you can challenge perceived wisdom as to how you should produce pages to convert people. There is no debate about which approach will work best for your site, as you just simply implement multiple approaches and see which actually perform best.

And then when you've found your best performer you make incremental versions of that to improve it even further.

I got the split run idea - thanks!

Forgive me though I am a slow learner...

So if I have a 20 pages website and this generates 1000 visitor a day organically... and I adjust to a long page version and my site is now 5 long pages and the traffic I have converts better so I implement... and because I lost 15 pages of content that have lots of links to them - my ranks for those pages will not drop, Is this correct?

Or will the causation of better conversions be lost traffic from lost ranks? In a PPC world that isn't an issue.

Or must I come up with 30,000 additional words first for these long pages so I can keep the organic ranks and traffic I have?

I understand better conversions are better but at the expense of earned exposure??

You experiment to see what works for you site.. there's no one size fits all.

Moz for example didn't concatenate multiple pages to create their sales page, they just increased a single page, most of the traffic to that page wouldbn't have come from the SERP's... however I bet that there was a call to action to get people to that page from the rest of the site that was added.

The thing to learn from that example is not to put all the content of your site on a single page.... it's to experiment with your sales process to find the most profitable route.

As I said, I don't concentrate on SEO for income growth anymore. So if traffic reduced but sales soared then thats what I'd go for!

In practice when CRO'ing pages it's usually quite straightforward to keep your ranks and traffic as long as you aren't a complete SEO numpty!

The real key thing with CRO is to experiment to see what works for your site, don't accepts perceived wisdom.. test it!

The Conversion Rate Experts did a talk at a conference I was at last week that maybe explains what's important in CRO better than I can, see it at:

http://www.conversion-rate-experts.com/conversion-killers/

And there is never a one-size-fits-all answer to conversions. Your traffic behavior will be different than people doing case studies and as long as you can keep the readers engaged throughout the long article/page. Multivariate testing will work to see what variations convert better.

You're right "one size doesn't fit"

http://www.youtube.com/watch?v=erdEZvOq6wo