Multi-Channel Attribution: Demystify Your Traffic & Optimize Your ROI

Multi-Channel Attribution: Demystify Your Traffic & Optimize Your ROI

This is a special guest post from Edin Šabanović, a senior CRO consultant at Objeqt.

Forget about social metrics like shares, comments, and “engagement.” Forget about advertising accolades and awards. Forget about open rates, clickthrough rates, and relevance scores.

The only way to measure the effectiveness of your marketing is to know (1) where your visitors come from and (2) what those visitors do once they arrive.

That means the secret to higher ROI lies buried in your analytics. Unfortunately, buried is the operative word … especially when it comes to multi-channel marketing that spans the billion dollar online-to-offline divide.

It’s one of the thorniest, most difficult issues ecommerce organizations face. Thankfully, there’s an answer.

Multi-channel made simple


Shopify Plus offers 16 native sales channels across social, Amazon, retail, wholesale, and more …

All without plugins or custom development.


Connect with an ecommerce expert today

A Crash Course on Google Analytics Attribution

Where: Source and Medium

Source and medium are the basic terms that describe channels in Google Analytics. For the purposes of conducting an omni-channel analysis, it is necessary to understand the way Google Analytics processes and tracks visit from these channels.

The terms “source” and “medium” refer to the individual channels visitors use to find and navigate to your website or app.

Google Analytics’ default grouping of various channels contains the following categories:

Source refers to a specific website or service that brought the visitor to your website, such as a tweet, search engine result, link in an email, etc. Medium refers to the category of the source: ie Social, Organic Search, PPC campaign, or referral from another website.

Top acquisition channels displayed in the Analytics Overview

Top acquisition channels displayed in the Analytics Overview

While Acquisition reports in Analytics give you a view of what sources and media bring visitors in and how many come from each source, you’ll remain none the wiser as to what those visitors actually do on your site. So you need to dig a little deeper.

You can use the “secondary dimension” function in most of the reports in Google Analytics to view how visitors coming from each channel behaved on the site. Viewing a landing page report using a secondary dimension of “source/medium” will reveal, for example, what specific services and categories brought visitors to that particular page.

However, using this method to estimate the contribution of a particular channel would be relatively inefficient. Fortunately, Google Analytics provides a special report called Multi-Channel Funnel, which tracks and reports how each channel performs relative to others.

What: Conversions

Multi-Channel Funnel reports in Google Analytics contain the details not only of the individual channels and medium that brought a visitor to the site but whether a particular visitor converted or not. Moreover, by tracking a visitor’s conversion path — “the sequences of interactions (i.e., clicks/referrals from channels) during the 90 days that led up to each conversion and transaction” — they can also tell you what goals, like time on site, number of pages viewed, etc., contributed to a purchase.

In the context of multi-channel analysis, this is the first step toward meaningful analysis. Multi-Channel Funnel reports contain the performance of the traffic channels we can explain — that is, the channels we know and can account for in terms of marketing activities.

Broadly, this includes …

  1. Organic search
  2. Referral traffic
  3. Direct traffic
  4. Paid search
  5. Custom campaigns

As long as your goals are properly configured, Google Analytics can create visualizations to gauge the relative importance of each channel. The overlapping areas represent the amount of interaction between various channels as well as identify the role each channel plays in the actual process of conversion.

Google Analytics offers insightful visualizations to simplify multi-channel attribution

Google Analytics offers insightful visualizations to simplify multi-channel attribution

To further analyze channel contribution, Google Analytics’ Assisted Conversion report “summarizes the roles and contributions of your channels.”

The following table — drawing again Google Analytics’ own data and nomenclature — represents each channel’s contribution with a number. Numbers below 1 signify that the channel generally converts visitors directly, while numbers above 1 indicate channels that “assist” in conversion.

Assisted Conversion is another useful report for analyzing multi-channel attribution

Assisted Conversion is another useful report for analyzing multi-channel attribution

The data contained in the Assisted Conversion report hinges on which attribution model you select… so let’s talk about different attribution models.

How: Attribution

How various channels contribute to online sales is the entire point of this analysis. To attribute a sale to a channel means to show the proportion of visitors this channel brings who convert along with the role those channels play overall.

Using attribution and attribution modeling in Google Analytics, it is possible to determine the relative value of the marketing channels in terms of conversions and ROI. The model you select greatly determines the outcome of the analysis and the value of each channel:

You can compare different attribution models within Google Analytics

You can compare different attribution models within Google Analytics

Deciding which model to use is well outside the scope of this article (which is why we wrote a whole separate article about it). Having said that, here’s a brief explanation of different models:

1. Last interaction

    The last interaction model gives full attribution credit to the visitor’s last interaction prior to conversion. All other activities of the visitor on your website are discounted.

    2. Last non-direct click 

    This is the default model of Google Analytics. It gives full attribution credit to the channel that precedes conversion unless that channel is direct. The general idea is to eliminate the unknowable “direct” from the equation.

    3. Last AdWords Click

    Full conversion credit goes to the visitor’s last AdWords click. This model is used to estimate the efficiency of individual ad campaigns and can be easily setup.

    4. First click 

    First click gives full attribution credit to the first click interaction. This model’s intention is to reveal channels that help establish brand awareness.

    5. Linear

      The linear interaction model gives equal credit to each channel. This model is not that useful, as it gives all channels the same amount of credit, regardless of their actual contribution.

      6. Time decay

        The time decay model gives most credit to the most recent interaction channel but still allows for earlier click to play a role. This is one of the most realistic default models.

        7. Position-based

          The position-based model gives interaction credit according to a channel’s position in the funnel. It is possible to customize the amount of credit given to each position.

          8. Custom

            A custom model can be adjusted manually according to the needs of the analyst. It is the best bet for coming up with a realistic model for conversion attribution, but requires extensive knowledge of the website’s business model.

            Remedying the Gaps of Multi-Channel Tracking & Analytics

            Using the above features, tracking and analyzing the vast majority of your online marketing spend (e.g., email, paid, or SEO) is easy. From there, you can concentrate your budget on the best-performing channels as well as optimize your channel content beyond the first click that brings visitors onsite.

            Instead, you’ll be able to test how that initial content converts over your entire funnel:

            The source/medium report uncovers the dynamics behind your multi-channel acquisitions

            The source/medium report uncovers the dynamics behind your multi-channel acquisitions

            The limitations occur when visitors reach your website by a direct medium. This can happen either …

            1. Online when someone clicks a link within a desktop application, secure website, document (.pdf, .docx, etc.), or direct messaging app (Skype, Viber, etc.)
            2. Offline when someone types the URL of your website in their browser as the result of traditional advertising like television, print, or non-linked promotions like Instagram

            The common denominator is sourcelessness. Direct visitors leave absolutely no information about how they got to your website or how they found it.

            To measure and optimize the efficiency of your marketing channels, you have to fill in those direct gaps.

            To remedy this unhappy situation, there is a way to give direct traffic some designation, enabling Analytics to show the channel that brought this traffic.

            (1) Filling the Online Attribution Gaps

            Urchin Tracking Module (UTM) parameters were devised to supplement the default mechanism used by Google Analytics.

            UTM parameters work by creating a unique “tag” you can add to any URL that contains tracking information about the channel, such as Source, Medium, Name, and Content. Creating a UTM parameter — either by using Google’s own Campaign URL Builder or another provider — is a must for any channel that will not be recognized by the default Google Analytics setting.

            Most commonly, this method is used to identify and track email campaigns, since many visitors will open their email from a client, which — devoid of a parameter to identify it — will be classified as direct.

            But, UTM parameters aren’t limited to email. They can be used to identify and track any digital source, from social campaigns, messaging apps, PDFs, offsite branded content, and even your own internal linking.

            Of course, if all your marketing efforts were confined to the online space, you’d have no problem tracking, attributing, and analyzing results.

            However, for high-volume ecommerce websites and businesses that combine an online presence with brick-and-mortar stores, multi-channel doesn’t just mean multiple online channels but offline channels as well.

            What do you do then?

            (2) Filling the Offline Attribution Gaps

            Offline channels present a unique set of challenges to online tracking tools.

            The first is the most obvious. If a visitor arrives at your website as the result of an offline marketing campaign, all the UTMs in the world won’t help you identify that channel.

            Second is the issue of engagement. For online sources, you can place links — calls to action — in particular, on-page locations, within offsite display ads or branded content, or at specific timestamps for video. Offline, no such luck.

            The third problem is visitors arriving at your website from offline sources have different expectations than those using online means. If you could detect who the visitors reaching your website using offline sources were, you’d have another opportunity to optimize.

            Let’s look at five ways to solve this issue:

            1. Provide a specific link in the commercial content

            For every offline campaign, create and promote a specific landing page. This isn’t just smart from a marketing standpoint so that the offline campaign leads directly to relevant onsite content, it also enables you to track the performance of offline campaigns. For longer URLs, use QR codes so prospects won’t have to manually type them in.

            2. Provide an easily identifiable and trackable incentive

              Providing a coupon code to prospects in print ads or within influencer marketing campaigns enables you to track and attribute each visit to that specific channel. Just be sure those codes are unique to each ad.

              3. Use pre-purchase or post-purchase surveys

              Attributing acquisitions and conversions is also possible (though less accurately) by using surveys. With this data, you can ballpark through representative samples how many customers discovered you from offline channels.

              4. Correlate by time or campaign duration

              By correlating traffic with the times an ad is aired — and isolating that influx from the regular direct traffic your website receives — spikes and conversions can be readily attributed.

              5. In-store displays

              In addition to multi-channel, numerous brands are now adopting an omni-channel approach to retail that includes in-store features like tablets located near cash registers where customers can fill out a survey, check catalogs, explore inventory, and so on. All those actions can also be tracked and assigned appropriate parameters.

              Removing the Multi-Channel Attribution Mystery

              Implementing online and offline tracking isn’t easy, and there are many obstacles that you’ll need to overcome to get meaningful results.

              The benefits far outweigh the difficulties.

              The ideal multi-channel marketing approach does more than just offer customers the ability to connect with you and order through their preferred medium — whether that’s onsite, via social, or through Amazon. Providing customers with this kind of experience and tracking how they behave, you can to create a customer journey that spans smoothly from online to off and back again.

              Remember, the only way to measure the effectiveness of your marketing is to know where your visitors come from and what those visitors do once they arrive. The secret to higher ROI lies buried in your Google Analytics … but it doesn’t have to stay buried.

              Sell anywhere, with just a few clicks

              Shopify Plus offers 16 native sales channels across social, Amazon, retail, wholesale, and more … all without plugins or custom development.


              Connect with an ecommerce expert today

              About the Author

              Edin Šabanović is a senior CRO consultant at Objeqt where he helps ecommerce stores improve their conversion rates through analytics, scientific research, and A/B testing. Edin is passionate about analytics and conversion rate optimization, but — for fun — he likes reading history books. Get in touch with Edin if you want someone to take care of your CRO efforts.