Step-by-step tutorials

Conversion optimization

Market discovery

1. Define a niche market

Some external sites, like facebook.com and linkedin.com, can bring different types of visitors to your site with some specific geographic, sociographic, and demographic features.

For example, use the Facebook ad manager to define a proper targeting to your product. Maybe target your competitors? Or perhaps technology early adopters?

Volument can track these niche markets separately so you can see how they generate traction and how they compare to the other segments.

2. Post a link

To track a niche market, you must add a special query string parameter to the link. This parameter is automatically generated when you add a new niche market to Volument. Few examples:

// Link to the front page. "m1" indicates a niche market
https://acme.org?m1

// link to a blog post. "m2" is another niche market
https://acme.org/blog/my-entry.html?m2

// using Google Analytics and Volument together
https://acme.org/?utm_source=linkedin&utm_campaign=my-campaign-1&m3

Volument places the visitors on the same analytical bucket based on the value on this parameter.

3. View results

You must wait for a few days to see how the visitors interact with the site, how they come back, and how they eventually convert. This takes a couple of days, depending on the amount of traffic.

Page optimization

1. Find out what to improve

The most effective way to optimize conversions is to continually fix your website bottlenecks because those are your “low hanging fruits” with the largest room for growth.

Volument gives you two perspectives to these bottlenecks:

Pages with the largest bounce count — Here's where the majority of your first-time visitors leave the site in just 7 seconds. There is something wrong with the content of these pages.

Pages with the largest bounce rate — These pages have the worst messaging. But the pages aren't possibly your biggest problems (yet), since they are not sorted by the bounce volume. Given more traffic, they would become your next leak on the bucket.

2. Find out how to improve

After deciding on the page to be fixed, you should check how people engage with the page, and it's viewports:

Are new visitors leaving on the first viewport? Fix your messaging and make sure the visitors understand the problem you are trying to solve.

Is the problem on a later viewport? Please check that out and think hard why people run away. Poor content? Complex layout? Are you selling too early?

Are people viewing just the landing page and not continuing to the next one? Missing links? Are links not visible? No good reason to proceed to the next page?

You should compare the page with the best one so you'll understand what works and what scares people away.

3. Publish a new version

With Volument, you don't need any extra steps on your workflow to push out new versions for A/B testing. You can use your existing deployment methods like GitHub pull requests to commit changes.

You tell Volument the time and description of the update, and you are good to go. People familiar with traditional A/B testing will find the following benefits:

  1. No need to change your working habits to optimize your website.
  2. No need for massive client-side libraries to slow down your pages and damage the first impression.
  3. No ugly content flashes from A to B to hamper the first impression. Sometimes you can see the original version, which is quickly replaced with different content.
  4. Anything can be tested: bug fixes, layout changes, design tweaks, and content edits. Change things individually or all at once.

Tip

Users of Optimizely can try to remove the massive JavaScript file (796 KB) and use Volument to A/B test the difference. How much more traction you get from the increased page load time?

4. View results

Once you've saved your new version to Volument, you'll start getting the results. The speed of the results depends on the type of A/B test:

  1. New A/B test. If the new version was pushed very recently, you must wait for enough visitors to appear. This usually takes a few days or just a couple of hours on a more busy site.
  2. Retroactive A/B test. Tell Volument about your change at a later time. In this case, you can get the results immediately in case there are enough samples.

Volument is an order of magnitude faster, reaching statistical significance than traditional A/B testing due to the following:

  1. You get tens of different engagement data points from every pageview. No need to wait for conversions at the end of the funnel that happen much more rarely. Typically one conversion per hundreds of pageviews.
  2. No need to collect data for the original version (A). The data already exists.
  3. You get exact engagement data for the tested page. The conversions never occur on a different page, and new- and return visitors are never mixed. Accurate data requires fewer samples to “smoothen the edges,“ and there is no need for A/A tests to “sanitize” the test environment.

Site optimization

1. Learn the problem

In site optimization you don't look for individual pages that are broken, but instead start from the current site version and look for general problems:

Low 3s stay rate? Too many (tracking) scripts to slow down the page load? Complex navigation? Content covered with a sales overlay or a large cookie consent dialog?

Low 15s stay rate? Is the product hard to understand? People cannot grasp the problem and how the product tries to solve it?

Low time engaged? Is the content boring? Hard to find more compelling content? Complex navigation? Complex header or footer?

Low retention? The content might be interesting, but not something worth a new visit? No re-engagement emails are sent for the newly registered users?

Low conversion rates? In most cases, the problem is on the above layers on the funnel: the product is not something people want. However, if you see high return rates, the problem is easy to fix: the call to actions are broken or simply missing.

2. Publish a new version

A unique feature of Volument is the ability to measure the success of bigger, site-wide changes. You can, for example, A/B test the following things:

  1. Global design changes
  2. Global content updates
  3. Changes on visitor onboarding
  4. New product launches
  5. New call-to-actions
  6. Changes in pricing

You can do radical experiments to see what works and what scares people away concretely. This is critical for startups, who must make radical pivots to survive. And all versions are saved for future reference.

3. View results

You must wait for a few days to see how the visitors interact with the site, how they come back, and how they eventually convert. This takes a couple of days, depending on the amount of traffic.