Research experiment – 10-200% increase by funneling visitors

website design - dan Hi all, the post that follows is the first results from a bunch of experiments we are doing to help us build our ‘Australian small business website design guidelines’. We want to learn from our own experiments (and other places) what works.I have to say the results were very encouraging. It also needs to be considered though that we only allowed 1 week before the changes and after the changes to compare the data. I have since been back in and looked at how the site is doing after 4 weeks in total and the figures aren’t quite as encouraging as below although they still represent a significant increase in every one of the items measured. In future we will try to give it 2 week either side of the changes for more comprehensive results.I hope you enjoy the post and please feel free to comment. I am more than happy to share any other stats people are interested in.

There are a number of assumptions web designers make about the optimum way to design homepages. Some of these are based on experience, some based on existing research from overseas. A lot are probably done for design reasons in some cases with no consideration for how it impacts on the site’s performance.

For this reason we have recently begun undertaking a series of research initiatives where we make some changes to one of our sites (in isolation) and we track the key measures of the site to see what impact it has.

The ultimate goal of the research we are doing is to develop a set of Australian small business website design guidelines based on our experience and our research onto “what works on the web”. Once complete, we will make this document available to the public free of charge.

The first in these research initiatives is testing the idea that the designer should take an active role in directing the visitors to where the business owner wants them to go and where they think the visitor wants to go. Rather than filling the homepage up with everything about a business or just a series of menus, we want to provide some obvious links to the common parts of the site so people don’t waste time trying to find what they are looking for and we are able to narrow down the people we really want to target.It’s called ‘funneling’ or using the homepage as a funnel as opposed to the common practice of cramming as much as possible into the homepage.

The results are very encouraging as you will see below.

The site and the changes

The site we have used for this first research experiment is this current site –

This is not our main company website (which is, it is a blog we have developed specifically to focus on website design topics for small business. However the site gets more traffic than our main site because of where it ranks in the search engines and we do get a lot of clients through it. So we get 2 types of visitors to the site, those interested in our small business website design blog posts (may not currently be interested in using us to build them a website) and people specifically looking for a website designer for a project they have.

This is how the site looked before the changes:

There were 2 menus, the top menu was for our services (designed to target people who are looking to have a site developed) and the bottom menu was for the blog (targeting visitors looking to read our small business website design blog posts). It was a lot for a user to take in and we felt we really weren’t doing our best to funnel the users to both (a) where we wanted them to go and (b) where we thought they wanted to go. We were getting a lot of hits on the website design category in the blog but we suspected this was a missed opportunity as at least some of these people were no doubt interested in our website design services and therefore our website design page.

This is how the site looked after the changes.

*Note we have since refined this design further so it’s a bit different to how it appears currently.

They are very simple changes. All we did was move the top menu to the top right out of the way and introduce 3 large modules representing where we want visitors to go and where we think visitors are going to want to go. Our main concern is making sure anyone who is interested in having us build a website for them doesn’t go down to the blog and get lost in all of the content.

Sample and conditions

The sample for this experiment was fairly small (as mentioned above), we allowed 1 week pre-changes and 1 week post-changes and the visits during each time frame were 300-400 (according to Google Analytics). We feel this is enough to give us a good indication of the results however acknowledge that it’s a fairly small number to draw too many conclusions especially on some measures like contact form completions where there was only a handful.

The measures

There are a few things we wanted to measure here. Firstly our main concern was ensuring people interested in our website design services are taken to our website design page instead of getting lost in the content of the blog so we wanted to measure the % of visitors who made it to this page. Secondly if we are trying to generate more business from the site and doing so by ensuring potential customers are taken to the most relevant page, we wanted to measure the ultimate goal of the site which is to get people to the contacts page and to complete the contact form. Finally we figured that the less visitors are confused and the more we push them to where they need to be, the more likely they are to stay on the site and the less likely to bounce to a different site straight away so we measured the bounce rate and the length of time on the site. So in summary the measures were:

  1. Visits to the web design services page
  2. Contact page hits
  3. Contact form completions
  4. Time on site
  5. Bounce rate

We measured these statistics for 1 week with the old design and 1 week with the new design.

The results

The results of these changes were really very encouraging. Every single one of the measures above were improved with the new design. Some by 10%, some by almost 200%. Here are the specific results for each measure.

Result Measure 1 – 180% improvement in visits to the web design page

Before the changes we were getting 3.05% of visitors to our web design page. After the changes we were able to increase this to 8.53%. This represented an increase of almost 180%. As this was the main purpose of the exercise these results were very encouraging. So at the very least we know we are getting significantly more people to a page that talks about our website design services.

Result measure 2 and 3 – 11% increase in visits to the contact page and 100% increase in contact form completions

The second measure was the % of visitors who made it to our contact page. In some ways the contact page is now a bit harder to find since it’s not sitting up the top by itself anymore. However since we figure we are keeping people on the site longer and giving them the content they want (and we want for them) we are more likely to succeed in getting potential customers to contact us – especially those visiting our website design page. And that indeed is the case if you look at the charts below, both the visits to the contact page and the contact form completions have risen significantly. The contact page hits rose from 3.82% of all visitors through to 4.25% and the form completions rose from .48% to .98%. Note this was only a difference of 1 completion so we can’t read too much into this given the small amount however it’s an increase nonetheless.

Results measure 4 – Users time on the site increased from 1:16 to 3:40

How long the average user spends on a site is an indication of how well you are engaging the user and how well you succeeded in helping them find what they wanted to find. Before our changes our average time on the site was 1 minute 16 seconds. After the change it increased to 3 minutes 40 seconds. This is a very significant increase and we will monitor this more over time to see where this ends up stabilising (note after 4 weeks post changes this was averaging 2 minutes).

Result measure 5 – Bounce rate reduced from 68.5% to 61.76%

The bounce rate tells you how many people left the site without clicking on anything (or in other words bounced as soon as they saw the homepage). People will bounce to a new site for a variety of reasons but obviously ideally you don’t want people to leave the site – you want them to continue to a relevant part of the site. Before our changes our bounce rate was 68.5% and after the changes it came down to 61.76% representing an improvement of 9.84%. This again was quite encouraging and the message from this is helping people find what they are looking for is a good way to keep people on your site.


We are quite excited about this first experiment particularly the fact that we choose a lot of things to measure and all of them resulted in pretty significant improvements. While the sample was fairly small we are confident that the results are significant enough to show that funneling your visitors in this way will result in improved website performance.

There is a lot more to do on this website and other sites we look after and a lot more guidelines to test. As always please feel free to comment on this post or suggest any future design changes we can test.

Tags: , , , , , , ,

About Alex

Alex Retzlaff is the owner of A Website Designer and Web Circle.

Follow us at:
Twitter @webcircle
Facebook /webcircle
Web Circle website

One Response to “Research experiment – 10-200% increase by funneling visitors”

  1. Sandra June 1, 2011 at 11:23 am #

    Hey, thanks for sharing this data. It makes sense that the new design would work better given the two audiences, but having data to prove it is always useful.

Leave a Reply