Technical details

Volument is an easy-to-use analytics software built around the classic idea of “actionable insights.“ Instead of just providing data for the sake of data, we deliver insights that help you create websites and content people love. We base our insights on engagement and statistical significance.

For less technical intro: Learn how Volument works

Content engagement

Volument delivers insights based on content engagement, a general term to express how your visitors gradually build interest before taking action. We illustrate engagement in the form of a “conversion funnel,“ which looks like this:

Awareness, interest, desire, and action

Key metrics

Volument takes advantage of the classic AIDA- funnel as follows:

Awareness – the content “captures attention” if visitors actively engage with the content for eight seconds or more. According to a study by Microsoft, you have no more than eight seconds to grab your visitors’ attention, and if you fail to do that, the visitor loses interest.

Interest — the content “builds interest” if visitors actively engage with the content for 15 seconds or more. Several studies conclude that 15 seconds is a critical threshold for deciding if a visitor eventually converts. Openweb calls this the “15-second rule”.

Desire – the content “engages” if visitors actively engage with the website for 60 seconds or more. Engaged visitors are more likely to take action. Volument uses this metric to measure the overall performance of your website, not the individual pages.

Action — the content “converts” if visitors react to any of the site's primary call-to-action, such as joining a mailing list or contacting sales.


Volument compares your content performance against benchmark values, a global average from other websites we measure. From “bad” to “great,“ you quickly see how your website performs.

Top performers

There are two kinds of well-performing content: interesting content (by the “15-second rule”) and converting content.

Interest is a more effective metric for good content than conversion rate because there is an order of magnitude more sample data available. All pages have readers, but only a few ones convert.

We mark each entry without statistical significance using a red dot. These entries have too little data and should not be used in decision-making.

Statistical significance

Volument requires at least 100 conversions on a content piece before you can trust the result. If there are less than 100 conversions, the outcome is meaningless and should not be used for insights. Conversions happen at all stages of the funnel: “became aware of the problem,“ “became interested,“ “took action,“ etc.

Most statisticians agree the minimum sample is 100. The good maximum sample size is around 10% of your audience, as long as this does not exceed 1000. For example, if there are 5000 visitors on a sample set, 10% would be 500 conversions. Sampling more than 1000 people won't add much to the accuracy given the extra time and money it would cost.

The sample size is heavily impacted by variance. If your variance is high, you must choose a number closer to the maximum. Conversely, you can go with fewer samples when the trend is stable. Content engagement trends will always look like this:

The variance is lower at the top of the funnel and higher at the bottom because more variables impact the metric's behavior.

For example, your visitors create a first impression of your website on the initial viewport of the landing page. The impact can be immediately seen on the top-of-the-funnel metrics when you change the title or add a new hero image. Conversion rate, however, is dependent on many more factors than just the title: a visitor can visit multiple pages over several days and visits. There are so many variables at play that some companies use extremely high sample sizes, even as high as 3,000–4,000

To defeat the impact of variance, Volument uses the following sample sizes for reliable results:

Confidence levels and corresponding sample sizes
Funnel stage Low Strong Very strong
Awareness 100 200 400
Interest 150 300 600
Conversions 400 800 1200


Volument uses drop rate to measure if a page fails to grab visitors’ attention. This metric utilizes the “8-second rule,“: i.e., you have around eight seconds to capture visitor attention. This metric helps you see what turns people off:

Note that the drop rate is different from GA's “bounce rate,“ which is the ratio of single-page visits, regardless of how much time the visitor engaged with the content.


Dimension is a descriptive attribute of the data: browsers, landing page, country, referring domains, and devices are default dimensions in Volument.

Unlike others, Volument pays more attention to content dimensions because the content is actionable. You can easily modify your content dimensions, but it's much harder to change your audience or market.

Volument is all about actionable insights that lead to increased engagement and conversions.

Content tagging

Volument uses content tagging to understand all the different content dimensions. It's a way to group similar content and give meaning to your pages beyond just the pathname. For example: grouping your pages with a similar subject helps you understand which topics work and which topics turn people away.

Content tags are provided with the HTML <meta> tags. Some tags, like the “subject,“ are standardized, and it's a good habit to provide them anyway. Some tags are dynamically determined, like the content length, and some are custom tags. Here's the full list of supported tags:


This is the value of the <meta name="subject"> tag. for example “Technology”


The type or function of the page given by the author. Example values: “home”, “article”, “podcast”, “webinar”, “index”, …


Value of the <meta name="target-audience"> tag. For example: “developers”, “designers”, “bloggers”, …

Actionable insights

Actionable insights are recommendations on what to do next, expressed in human language, and they are the holy grail of web analytics. This is the actual value of Volument: you'll learn what to do to make your content more engaging and gain more conversions as a result.

We only list statistically significant insights. For this reason, the list is initially empty, and new insights appear as they have enough samples for confident results.

There are three types of insights in Volument: warnings, suggestions, and status notes. Some examples:

Warnings, suggestions, and status notes


These are usually sudden drops in engagement or conversion rates. For example, if your “Contact us”- page suddenly starts turning people away more than usual.

Another important category of warnings are leaks. These are the content pieces that drop visitors the most. We measure this by drop volume, not by drop rate. These content pieces are your “low hanging fruits” where you can make the biggest impact with the lowest possible effort. We mark significant leaks, with over 20% of your total audience, with an “urgent” tag.


Suggestions advise you to switch from one thing to another. For example, we might suggest you should write more about “Technology” topics than “Marketing” topics, or advice you to stop playing around with “webinars” and advice to use “explainer videos” instead.

Status notes

Notes are non-critical issues, such as a reminder that you haven't published anything new for three months or a note about a page without any traffic. We also tell you about good things, for example, when your engagement or conversion rates suddenly go up or your website has a significant traffic spike.

{"style":"/docs/doc","desc":"Volument is an easy-to-use analytics software built around the classic idea of \"actionable insights.\" Instead of just providing data for the sake of data, we deliver insights that help you create websites and content people love. We base our insights on *engagement* and *statistical significance*.","title":"Technical details","url":"/how-it-works/technical-details","key":"technical-details","created":"2022-04-29T11:18:29.987Z","modified":"2022-11-08T06:40:37.533Z","createdISO":"2022-04-29","modifiedISO":"2022-11-08"}