The challenge and promise of digital campaign tracking

Have a look at this image. Sound familiar? Web analytics has held out the elusive promise of being a set-it-and-forget-it kind of thing. "Set up your reports, and the data will fill itself in. That promise has largely held true - for every part of web analytics except Marketing. That's because with marketing, the web page you have today isn't the one you had yesterday.

There's constant change: new information, new deals, new parameters. What everyone wants is a system that runs itself. Otherwise, as the figure shows, you spend all your time making sure the reporting is right. Spending time on data correction takes time away from the analysis that will really help the company. It's a necessary evil. Wouldn't it be great if we could get marketing data to the same set-and-forget kind of place as the rest of our web analytics?

Multiple digital channels, manual data handling

Our primary challenges come from two sources: integrating data from multiple data streams, and handling data manually. More and more, tracking codes are being generated automatically by the ad platforms a marketer works with (DoubleClick Manager, Adwords, other paid search mechanisms). But those codes are unlikely to follow a brand's parameters for the data they want to gather. The tail wags the dog when you engage with these systems: they build tracking codes their own way. Adobe can track the data, but it can't classify for you - it isn't set up to do so. Someone (probably you) will have to go in and sort through it all. The promise of these systems is that they will eliminate the manual generation of tracking codes, when in fact they just move the manual data-handling problem from the front end to the back end. The fact that codes are being created on the fly means you're always playing catch up.

Complex campaign targeting and naming conventions

Google Analytics has tried to simplify the process by reducing the amount of choice. If you're limited to five measurement variables (paid search, source, medium, campaign name and content) that should make this better, right? Except that a lot of companies can't really fit themselves into those five parameters. One specific GA user we deal with uses a mini-formula for the campaign name. Their "simple" campaign names included: date, brand, ad type, pitch, audience, and strategic objective. They do the same thing with the other fields. Trying to measure 25 different values to stay on top of their various campaigns ends up twisting things in the opposite direction from the simplicity GA promises. It's disheartening to watch companies contort themselves to get more and more info from those five variables, to the point where they're disconnected from the original metric.

Current Solution: Excel spreadsheets

To get around these problems, most companies still use spreadsheets. The upside is that you can use formulas and lookup tables to create consistent tracking codes, regardless of which analytics tool or ad platform you use. The introduction of sheets that can be shared across an organization has helped a lot (solving the problem of many documents in siloes across the organization). However, the simpler you try to make things for marketers, the more hidden complexity you need to add to your tracking codes to keep the data up and running. That spreadsheet is valuable, but it dies with you. When you leave the organization, all the hard work you put into that document to make it simple for users makes it impossible for the new person to understand what you were doing. This is a terrible situation for businesses that want data continuity or any year-over-year picture of macro trends for the business.

Numerous fail points in the process

What we need is not just more but better automation, including automated communication from the ad platforms that create codes for you to put into your system. Your system should then integrate those link parameters in such a way that it communicates to your analytics tool the pieces of data that are important for you. If we can get ahead of our data flow with automation, we won't constantly be playing catch up. Catch up means there's always a gap. Instead of spending the first 24 hours of a campaign seeing if you're capturing any data, you could be assessing if the campaign is achieving its objectives and whether the investment is a good one.

People sometimes say "Why worry, Adobe works retroactively. You can always go in and clean things up." Maybe. In the meantime, there's confusion and a business cost to making decisions without accurate data. The confusion and the running blind contribute to a credibility gap for analysts Having the data validated before it goes live gives you the Speed to Insight that organizations are seeking.

Part 2 - Best Practices for Digital Marketers

Everybody is busy with more campaigns, more targeting, more tools - a deluge of reporting data. It's well known that digital marketers are swimming in data. With only so many hours in the day and digital marketers often struggling to build-up their competency with analytics itself, addressing data quality issues falls way down on the priority list. Here are some areas you can focus on to change that.

Connection Rate

What does not get measured ends up not getting managed. One of the most useful metrics about the fidelity (faithfulness) of digital marketing campaign data is the connection rate. A low connection rate means that whatever engagement metrics are being measured (time on site, bounce rate, page views per visit) are misleading or low data fidelity. A high connection rate means that all derivative metrics are meaningful or high data fidelity.

Sometimes called the click-to-visit rate or match rate, this one metric can help stakeholders understand where their technology or more likely processes have gotten fouled up. Simply put this is a measurement of front end click data, e.g. paid search or display media counts and relevant visits on the destination Web site. Clearly, these numbers can never be 100% as they represent different yet related measures of traffic. However, anything below 85% is worth a closer look. If your service providers are not offering you a dashboard on analytics data fidelity, or at least periodic updates, start your data improvement effort here.

Mapping Tables

The best practice in the industry today is to use a single and preferably context-less campaign tracking parameter. The reason is that multi-parameter methods that are popular with Google Analytics default functionality create a very troublesome and error-prone workbook process. Derived data from the typical 3-5 name-value pairs is not enough meta information about a campaign. In addition, this method uses literal names in the value parameter that can reveal to competitors what is going on and how it is being tracked. Keeping all of these values in the cloud in a centralized database ought to be our gold standard as this enables downstream CRM databases to have consistent attribution reporting. Using Adobe SAINT or Google Analytics custom dimension-widening is best.

The Workbook-in-the-cloud

With the dynamic nature of digital marketing unlikely to change any time soon and the number of campaigns, vendors and stakeholders on the rise, having a single source of campaign tracking truth is essential. The promise of digital marketing is continual optimization. Unfortunately, that is also the greatest challenge: success is based on improving current performance. When analytics reporting is not reliable, when it has to be constantly qualified or requires many manual adjustments, this can lead to a no-decision scenario and worse: a lack of confidence in the platforms and teams running them.

The benefits of a unified cloud-based solution are many. With all agencies, internal teams and analytics stakeholders sharing one platform, the synergies are considerable. Syntax-checking can be programmed in to avoid simple typos. Enterprise rules and changes can be made once, and multiple teams can derive the benefit immediately. Validation of landing page URLS can be automated. Last, onboarding disparate teams can be routinized since the methods and rules are no longer siloed by team.

The road to oblivion is paved with good intentions - you need to Do It

Everybody says they want better data, but corralling the key stakeholders and measuring the extent of the problems can be daunting. It won't just happen on its own. Savvy digital marketers and analysts should get control over digital marketing data before it's too late. Using best-in-class operation campaign tracking tools and processes (like RASCI) can solve this thorny problem once and for all.

Authors