Set up funnel steps

On the Settings tab, along with setting up model parameters and channels that should not get value assigned, you can adjust the conversion funnel according to your project’s needs.

Basic settings

To manage a funnel step, use the context menu next to each funnel step:

When you add a new funnel step, you may choose its placement in the funnel: before or after the selected step. To do it, select the relevant action from the menu. Next, define this step conditions in the opened pop-up window:

• Parameter. Select a parameter that will represent this funnel step, for instance, visits to a particular website page.
  The list of the available parameters depends on the data source type you’ve selected at the model creation step. Learn more from the table below.
  

• Condition. Define a specific condition for the selected parameter.
  For example, to filter out user visits to the catalog page only, define the step conditions as: “Page + exactly matches + catalog”. Or you can set up the condition: “Hostname + doesn’t contain one of + hostname1, hostname1” to exclude any data from the given hostnames.
  To create a more complex condition, use either the “matches regex” or “doesn’t match regex” options. The regular expression is useful to search a word or a phrase from the list, a word with special characters or spellings, an exact phrase or the alphanumeric characters, etc. Also, you can apply the regex to filter out any email address within a preset domain or any IP address from the range of addresses.

• Value. Specify an exact value for the selected parameter, or provide a regular expression in format RE2 that defines this parameter value.

Advanced settings

You can also create more flexible definitions for your parameter by applying the advanced settings.

• First-level conditions. Use the option AND to make your step match several conditions simultaneously. The option OR allows creating a step definition where any of the contained conditions must match.

• Second-level conditions.  Allows creating nested conditions for the first-level ones to make them even more specific.

Once you complete the required fields in the pop-up menu, click Add step.

TipAt any time, you can edit a funnel step settings using the Edit action from the context menu. To save the edits, click Apply changes.

Conversion step and model calculation run

Once you define the interim funnel steps, set up the conversion step. Learn more from these instructions.

Next, click Preview funnel to ensure that all the settings are correct. Once you verify the settings, click Save funnel steps and settings, then run the model calculation.

Funnel metrics

Each step has four parameters: Step conversion rate, Probability, Value, Step abandonment rate.

Step conversion rate (blue chart) is the percentage of sessions in which a user had visited the step. For example, the “Checkout” step’s 3,7% conversion means that 3,7% visits included visiting the checkout page.

Probability (grey chart) shows the percentage of users who came to this step from the previous one in the funnel. For example, the “Checkout” step’s 62,2% probability means that 62,2% of “Add to cart” actions led to the checkout.

The grey arrows that appear when hovering over the chart show the probability of moving from the current step to the others.

While calculating the probabilities, BI Attribution segments the data breaking it down by user types, regions, and device types. The segmented data is available in the Results tab.

The Visit step probability. If you have a visit as the first step, then its probability equals 100% and value 0%. To assign value to the channels targeted at visitor acquisition, you can set “Non-bounce visit” as the first step. Its probability is equal to the share of no-bounce sessions.

Data is segmented during probabilities calculation.

During the probabilities' calculation for each step, the data is grouped by customer type, geo regions, and previous steps. The probabilities of how users go through the funnel significantly depend on these parameters.

Customer type — new or returned — is defined as whether a visitor made a purchase during the last year or not.

Device category — desktop only, tablet only, mobile only, cross-device — is defined by the category of the device a visitor has been using during the last year.

Geo region is defined by the Google Analytics session dimension. 

Value (green chart) shows the conversion value share. The value is being determined as follows: first, we calculate the step’s weight — the smaller the probability, the greater the weight. Then, the step’s weight is being divided by the sum of all steps.

Simply put, the more difficult for the visitors to get past this step the more value this step gets. 

Here's an example of value calculation (the same table in Google Sheets):

If you see an undefined value (NaN%) for a funnel step, read this article.