Direct Traffic

Rockerbox will define a conversion as being driven by “Direct Traffic” if there were no other marketing touchpoints that led a user to convert.

Example 1:

Path to conversion is Search->Facebook->Conversion. This would be reported out as:

  1. Search
  2. Facebook
  3. Conversion

In the above scenario, conversion credit would only be distributed between Search and Facebook.

Example 2:

The Path to conversion is Direct->Conversion. This would be reported out as:

  1. Direct

In the above scenario, conversion credit would be assigned 100% to the direct touchpoint.

Applying Attribution Credit to Direct Traffic

Example 1:

Path to conversion is Search->Facebook->Conversion. Conversion credit would be allocated as follows:

  • First Touch: Search receives 100% of conversion credit
  • Last Touch: Facebook receives 100% of conversion credit
  • Even Weight: Search and Facebook each receives 50% conversion credit respectively
  • Statistical: Search and Facebook get X% and Y% of conversion credit, based on the model. X+Y=100%

Example 2:

Path to conversion is Direct->Conversion. Direct would get 100% of the conversion credit regardless of attribution methodology. This means that when analyzing Direct as a channel, the result is the same regardless of attribution methodology as Direct only includes users that engaged with 0 non-direct marketing touchpoints.

Reducing Direct %

Reducing Direct Traffic gives you more insights into your marketing impact on user actions. You can use the following tactics to reduce Direct for your conversions.

  1. Map your existing Direct URLs
  2. Add more Marketing data
  3. Add more user identification data

A) Map Eligible Traffic

A potential Marketing Event will be treated as Direct if either of the following are true:

  • If doesn’t meet any of your existing mapping rules in Rockerbox
  • It meets a rule set in Channels > Exclusions

Review all URLs that are classified as Direct and see if they are classified correctly. It’s quite common for there to be URLs that should be mapped to a channel.

Less common are URLs that are excluded as marketing events but should actually be included. For example, Rockerbox has default Channels > Exclusions for many payment providers.

This is so that a checkout flow that redirects to a payment provider's website doesn’t count as a marketing event. However, you may have a partnership with a payment provider where they refer traffic back to your site. You would not want to exclude those referrals.

B) Add more Marketing data

Your customers might have heard about you from a channel that isn’t tracked through your current setup. In that case, adding more marketing data will reduce Direct. Examples include:

C) Add more User data

Passing user data to Rockerbox more frequently can help identify users. These tactics include:

Attribution Type

Since Direct conversions don’t have any marketing touchpoints assigned to them, changing the Attribution Type between First Touch, Last Touch, Even Weight, or Multi-Touch will never impact the number of conversions attributed to Direct.

FAQ

Why did my direct percentage increase?

  • There is an absence or interruption of a key user identifier in the conversion data Rockerbox receives - this prevents us from completing identity resolution between conversions received and marketing events.
  • There are specific contributors depending on how the conversion is being captured - batch, pixel, webhook - see table below for specific considerations.
  • Marketing events are missing tracking.
  • Marketing events have tracking but Rockerbox does not have mapping set up (most commonly for UTM based tracking).

Conversion Capture Method

Direct % Contributing Factors

Batch

1. Are we receiving the correct user identifier on conversion data 100% of the time?

2. Is our All Pages pixel firing on every page of your site in a timely manner, allowing us to identify users?

3. If Shopify, did you complete the Shopify Checkout Extensibility implementation? This is relevant only if you migrated to Shopify Checkout Extensibility.

Pixel

1. Is the Rockerbox All Pages pixel firing immediately with the Rockerbox UID on every page?

2. Is there any delay between site load and the All Pages pixel or conversion pixels firing? For example, is the "conv.purchase" pixel firing before the email is available or so delayed that the user has navigated away from the page?

Webhook

1. Is our All Pages pixel firing on every page of your site in a timely manner, allowing us to identify users for matchback when we receive webhook conversion data?

2. Did anything change with the persistent user identifer sent in webhooks? For example, did the name of the field change or did the value sent in the ID field change?

How did we do?