5 Steps to Identify Waste in Your Digital Ad Spend Using Google Analytics
May 03, 2017 Mark Bertrand Data & Analytics
Implementing a digital advertising strategy with these five steps will identify the effectiveness of your digital advertising spend. One of the most convincing reasons to adopt a digital advertising strategy is that it provides the ability to track every advertisement from start to finish in near real-time results. In the past few years, this benefit has been enhanced even further through the application of programmatic ad targeting. Despite these alluring benefits over traditional advertising venues, there is still a shared and often nagging problem -- that of identifying which channel is truly driving positive ROI performance. If you’re feeling as if your digital advertising budget is performing much like a stack of $100 bills sitting on an open windowsill during a windstorm, and if you’re using
Google Analytics to track revenue or lead generation performance, then the five steps defined in this article will help you identify where your advertising campaigns are wasting your advertising budget.
The five steps you need are:
2) Goal Value
4) Google Tag Manager
5) Attribution Models
Our Analytics department at LaneTerralever incorporates a minimum of five views within the Google Analytics properties when clients require budget analytics. These five views not only help to prevent data corruption and a loss of data should something go wrong, but they also prevent data sampling. For example, if you were analyzing the campaigns from your Google Paid Search channel and wanted to add a secondary dimension to the campaign analysis so you could see the actual keywords, when you add that secondary dimension, Google Analytics automatically begins using data sampling. The tables and graphs in the display are now representing only a small percentage (sample) of the total traffic your campaigns actually received. The reason Google Analytics (GA) uses sampling is to enhance performance. It speeds things up so that when you enter the secondary dimension and click to see the results, it takes just a fraction of a second to see the results. The alternative to sampling is to have your GA developer define a unique view for critical analysis you plan to use frequently. Going back to the example above, using Google Paid Search as a unique view, you will avoid sampling because the historical data collected in the view is already filtered to just what you need to analyze.
The five standard views used for budget analytics include:
1) Raw Data – this view is the backup for recovering, should the need arise.
2) The Main View – this is used for all of the standard goals, filters, and where access is provided to others.
3) Vendors – this view is used to provide third-party vendors with access.
4) Test – this view is where new filters, alerts, and changes to existing settings can be verified before adding to the main view.
5) Attributions – this is where we place macro conversion goals and funnels.
The Attributions View is key to defining where the advertising budget is being wasted. In this view, only the goals that are key to revenue growth are set up. The obvious goals to use would include purchases. However, for many businesses, the macro conversion may be phone calls, registrations including creating a user profile, or signing up to attend a webinar. Whatever the macro conversion for your business, the reason for including only macro goals in this unique view is to prevent conversion rate and conversion value contamination. In this Attributions View, you don’t want those “progress” KPI goals contaminating the “achievement” KPI goals.
While Google Analytics is a great data management platform for tracking your website performance against metrics such as pageviews, bounce rate, time per visit, and an abundance of other click-tracking metrics, these standard tracking metrics do not tell you how well your business is doing. There are over 100 customizable settings within GA, and if you don’t set them up from the start, you’re losing valuable information each and every day. Setting your analytics property(s) up completely and correctly is even more critical when you want to discover where your advertising budget is being wasted. Defining the achievement goals is a good start, but assigning them without a value is sort of like making a cherry pie but forgetting the crust. The real value of the system is truly lost. We need to break down, campaign by campaign, source by source, where the value is coming from, and where it is not coming from. The only way to do that is to set a value for every goal.
Even if you need to estimate or use an average for the value, you’ll be able to see the trends over time for channel and campaign performance. Trends based on estimates are just as impactful as actual revenue for measuring digital advertising strategy performance. Using goal values is not limited to just ecommerce applications. For example, if the macro conversion offered on your website produces leads, and the leads are phone calls, we can back into the per-call-value. Let’s say that a sale is worth $10,000 and the average sales person in the call center is closing 1 in every 20 calls. By mathematical division, we can estimate each call is worth $10,000 / 20 = $500. When the GA developer is setting up the Phone Call conversion goal, you’ll instruct them to set each goal to be worth $500.
There are more insights available beyond just identifying where your advertising budget is being wasted. When you set a dollar value to your goals, even in those other views where you’re tracking micro conversions, it provides an entirely different set of analysis capabilities in nearly all of the standard tables and graphs in the data management platform (DMP). You’ll quickly change the standard click-tracking metrics into story-telling – insights that identify opportunities for conversion rate optimization.
Bonus: Making your goal values even more accurate requires an additional calculation.
Rather than using the estimates or average value of a lead or sale, use the lifetime value (LTV) of a customer.
One last consideration to setting goal values before we move on to Step 3. Goals are like most everything else that needs to be set up in GA. These values are not set-it-and-forget-it; they need to be adjusted periodically. The most common reasons for periodically changing goal values include seasonal cycles, economic changes, and knowledge you gain over time for setting and estimating the goal value.
There are eight defined channels in the standard Google Analytics DMP. These standard channels include:
6) Paid Search
You can target these channels to identify where the advertising budget is being wasted. However, you’re likely going to discover that the traffic recorded and shown in your DMP is contaminated. For example, if your GA developer hasn’t already configured the channel grouping, you’re going to discover that Facebook traffic isn’t showing up in the Social channel where it belongs. It’s in the Referral channel instead. Most of the email clients don’t get picked up in the Email channel either, but are instead getting listed as Direct. Even more elusive is the traffic from retargeting campaigns and remarketing campaigns. That’s because the order of the channels plays a significant role. As you can see from the list of channels above, the standard order of channels in GA places the Display channel in the last position -- number 8. If your advertiser is using Google or Bing (or both display networks) to serve the retargeting and remarketing display ads, then all of the traffic from those ads is getting captured and shown in number 6, Paid Search. Those are just a few of the easiest to identify and correct. An experienced GA developer will know where to look and how to correct these channel groupings.
While having eight channels to discover where your digital advertising budget is getting wasted may seem adequate, these eight are just the aggregates. Your GA developer can create additional channels for drilling down deeper, channels that are unique for your business needs. You can then define the specific campaigns and the mediums for each channel, and when your advertising links are using a well-defined UTM strategy, there isn’t any advertising effort that can escape your analysis. Not even the offline advertising strategy. Some of the most often used GA channel customizations include creating additional channels for Traditional or Offline, Retargeting, Remarketing, Local Directories, and Affiliates.
Channels can also be created and developed to accurately capture and show offline advertising campaigns. When these campaigns include consumer opportunities to use barcodes, scannable coupons, vanity URLs, and hashtags, you can discover their value using GA. Worth mentioning also is cross-domain, where your shopping cart or cash register is on a different domain. All of these require a professionally trained Google Tag Manager (GTM) and GA developer. Speaking of GTM…
Anything that a consumer does on your website is trackable, including page scrolling, hovering over images, starting and stopping videos, starting to complete a form and then abandoning before completion, etc. While GA automatically tracks clicks and pageviews, you use Google Tag Manager to set up tracking for everything the consumer does, but only to the level it makes sense for your business and your digital strategy. The micro and macro conversions are first developed as events and triggers (funnels) in GTM, so that later in GA you can use those events to set up the conversions.
The term “attribution” in a digital advertising strategy refers to applying a mathematical model that credits traffic sources for a desirable behavior, such as sessions and conversions. Consumers will visit a website multiple times via multiple sources, and there are many models you could use to credit those sources. For example, you could give all of the credit to the first source, or all to the last source, or divide the credit evenly between all sources, or even use weighted credit depending on the position.
Sessions are automatically credited to the source in GA. All of the standard reports in the DMP use “last-click attribution” to report these traffic sources. That means whichever source the consumer used to reach the site this time (last click) is the source that receives credit for any conversions in the session. More correctly stated, the current source is attributed credit for the conversion.
There is one exception to this rule, and that is Direct. When the last source is direct, GA will use the previous source for the attribution. Since Direct is truly the absence of knowing where the consumer came from, then we don’t want to overwrite what is known (the previous source) with nothing. By default, GA recalls the consumer source for six months. The timeout is customizable from the GA Admin setting of the property under session settings.
There is a series of reports available in the GA DMP within the Conversions, Multi-Channel Funnels reports where you can see the source for every session over a period of time before the conversion. As stated above, this “last-click except direct” rule is true for all the GA standard reports. In this same section of GA, Conversions is also an attribution modeling tool. These models allow analysis for different attribution rules such as first click, last click including direct, weighted credit for all sources, and you can define your own models as well. If you have followed the four steps above and in combination with these attribution models, you can easily identify which channels are wasting your advertising budget.
Need help getting started or looking for more information on attribution modeling? Head over to our website to learn how we can assist.