Multi-Channel Funnels help you understand how your campaigns contributed to conversions. Picture a funnel that leads users from a wide variety of locations on the web — all the places where your ads were displayed — to the pages where you've set up Floodlight conversion tags. The reports in the Multi-Channel Funnels section give you insight into how users follow that path. Because Multi-Channel Funnels reports focus on the interactions that led up to conversions, they can give you insight into the role that your Campaign Manager channels — or specific ads, site placements, creatives, and click trackers — played in your conversions.
Channels are different media types within Campaign Manager. The default channels available in Multi-Channel Funnels reports are display ads, Search Ads 360 ads, Rich Media ads, and click trackers. You can also define your own custom channel groupings.
The metrics within these reports are generated from conversion paths, which are defined as the sequences of clicks and ad impressions that led to that particular conversion. The reports only include ad interactions that were part of a converting path. For example, if a visitor saw or clicked one of your display ads but never returned to complete a conversion, that display ad is not included in your Multi-Channel Funnels reports. You can also create custom conversion segments to analyze specific conversion paths.
View Multi-Channel Funnels reports
To view Multi-Channel Funnels reports, click the Attribution tab at the top of the UI, then select the type of report you want to view from the left-hand navigation section.Types of Multi-Channel Funnels reports
The following types of reports are available:
Conversions: The most basic reporting type that you can use as a baseline against which to compare other conversion data and report types.
Assisted Conversions: Learn how and when different channels had an impact on conversion. You can also see which channels belonged to an upper or lower funnel.
First Interaction: See which channels, campaigns, or sites are creating leads and are the first interactions in a path.
Top conversion paths: Learn the routes your customers take to a conversion.
Time Lag: View the amount of time to conversion after the first interaction occurred.
Path Length: View the number of interactions before a conversion.
Attribution Modeling Tool (AMT): Compare data using several different attribution models (such as last click or weighted) or build your own custom report.
To run a Multi-Channel Funnels report, you need to have the following permissions enabled (Learn more about user permissions).
At the user role level:
- View Conversions
- View Click-through Conversions
- View View-through Conversions
At the user profile level:
- Any advertiser filter
- Any campaign filter
- Any site filter
Exports are available for all types of Multi-Channel Funnels reports. Click the Export button, then select one of the following options:
CSV (comma-separated values): Each row of data is on a separate line, and each data field is separated by a comma.
TSV (tab-separated values): A simple text format, where each row of data is on a separate line and each data field is separated by a tab.
TSV for Excel: Similar in format to TSV, but encoded to be more compatible with Microsoft Excel.
The report will automatically begin downloading. The data included in the export will match your current view on the “Attribution” tab (that is, it will contain the same metrics and dimensions that you have applied to your report).
With the exception of the Attribution Modeling Tool (AMT), all Multi-Channel Funnels reports are based on the lookback windows that you set in Campaign Manager. For example, if you've set an impression lookback window of three days, only impressions within the three days prior to the conversion are counted as part of the path. Older impressions aren't counted. We use lookback windows to make sure that the data you see is fresh and useful.
Both clicks and impressions are used. If your lookback window for clicks is longer than the lookback window for impressions, it's possible that an interaction could be counted based on a click. In cases where both an impression and a corresponding click are within the lookback window, both interactions are counted.
If you’re using the AMT, you can manually configure your own lookback window. Learn more
Data samplingWhat does “sampling” mean?
Sampling means that there is too much data in a report for the UI to process all of it and load quickly, so we look at a percentage of traffic and extrapolate it to 100%. However, sampling is imprecise, and the numbers provided are not intended to match your actual conversion data. You can use the results to analyze overall trends and patterns in your conversion performance data.
Where sampling occurs, a message appears in the UI indicating the number of conversions used for the sample, along with the percentage of the total conversions this represents. For example: “This report is based on 1,000,000 conversions (8.86% of conversions).”
Sampling in Multi-Channel Funnels (MCF) happens whenever you’ve selected a combination of reporting criteria that would result in more than a million conversions being displayed. For example, if you’re reporting over a large date range, or your Floodlight configuration contains a large number of conversions, then sampling is likely to occur. Reports containing conversion segments will also use sampling.
The Attribution Modeling Tool (AMT) nearly always samples data for large networks. The purpose of the AMT is to allow users to build and apply attribution models in real time, so it’s not possible to pre-compute the values that would be needed to give precise values in real time. In other reports, such as Top Paths or Assisted Conversions, often the Basic Channel Groupings data won’t be sampled, but data for other primary dimensions (such as placement or site) will be.
Sampling occurs in Multi-Channel Funnels whenever you are looking at a Floodlight configuration with more than a million conversions. For the purposes of sampling, the number of individual conversions in a Floodlight configuration is less important than the aggregate number of conversions. For example, say that you’re running a report on several activities that only have around 5,000 conversions each. If they are within a Floodlight configuration that contains 1,000,000 (or more) conversions, we’ll still sample the data as a result of the size of the Floodlight configuration as a whole.
When sampling, every request generates a new sample with different numbers. If you refresh the page or run an export, the data will be different each time.
Yes. You will get a different result every time you load the page, since sampling happens differently every time, even with the same set of data.
Since sampling occurs independently for each column in your report, and sampled data is only representational, sampled columns will not add up.