A Framework for Digital Analytics Process

Digital Analytics process can be used to accomplish many things. Yet, in it’s most valuable form, process should be viewed as a means to familiarize business users with data that is potentially available to them and to create efficiency around how that data is collected, analyzed, and provided back to the business.

Most organizations have organic processes that grew out of necessity, but in my experience few have developed formal process for taking in analytics requests, for data quality management, or for new tagging requests. While each of these activities usually happens at organizations today, they are largely handled through ad hoc processes that fail to provide consistency or efficient delivery. As such, Analytics Demystified recommends that companies implement a process framework that will address each of these critical components.

Note that the introduction of a new process into a business environment requires a change in habits and routines. While our process recommendations seek to minimize disruption to everyday operations, some new ways of collaborating will be required. Analytics Demystified’s recommended processes are designed to be minimally invasive, but we recognize that change management may be required to introduce new process to the business and to illustrate the business benefits of using process to expedite analytics.

Digital Analytics New Request Tagging & QA Process

This process is designed using a Scrum methodology, which can easily fit within most companies development cycles. At the conceptual level, the Analytics Tagging & QA Process provides a method for business users to communicate their data needs, which are then used to: 1) Define requirements, 2) Create a Solution Design, 3) Develop Analytics Code, 4) Conduct QA, and 5) Launch new tracking functionality (see diagram below).

Tagging & QA Process — Starting Point:

The tagging and QA Process is one that is typically used by organizations multiple times throughout website redesigns, feature/function improvements, and general updates. It is intended to be a scalable process so that it can be used for all future feature and development projects that require analytics as well as digital analytics analysis requests.

The starting point for this process includes a “Digital Analytics Brief” that will be used to identify goals, measurement objectives, and specific elements that need to be tracked with analytics. We recommend using a simple Word, Excel or Google Doc document to capture information such as: Requestor, Request Date, Due Date, Priority: (Low, Medium, High), Overview: (brief description of request), Primary Objective: (What are you trying to achieve?), Desired Outcome: (How do we know if we’re successful?), and Additional Comments. Using a brief will force business users to think through what they’re asking for and to clearly define the objectives and desired outcomes. These two components are critical to determining success factors and formulating KPIs.  

A Digital Analytics Brief can be expanded over time or developed as an online questionnaire that feeds a centralized management tool as companies increase their sophistication with the Tagging & QA Process. Yet, either simplistic or automated, using this a Brief format as the first step in the data collection process will enable the digital analytics team to assign resources to projects and prioritize them accordingly. This will also serve to get business users accustomed to thinking about tracking and analytics early in their development projects to ensure tagging will be incorporated into development cycles.

Step 1: Defining Business Requirements

With the Digital Analytics Brief in hand, the business analyst should have pertinent information necessary to begin the process of defining  business requirements. Depending on the scope of the project, this part of the process should take between one and five hours to complete with the Digital Analyst leading the effort and stakeholders collaborating with details. Demystified recommends using a template for collecting business requirements that captures each requirement as a business question. (See Bulletproof Business Requirements for more details).

One of the things that we’ve learned in our years of experience working with digital analytics, is that business-users are rarely able to articulate their analytics requirements in a manner that can be easily translated into measuring website effectiveness. Simply asking these users what data they need leads to insufficient information and gaps in most web analytics deployments. As such, Analytics Demystified developed a process designed to gather information necessary to consistently evaluate the effectiveness of our clients fixed web, mobile sites, mobile apps and other digital assets.

By using a similar process, you too can effectively identify requirements and document them using a format ready for translation into a Solution Design document.

Example Business Requirements Documentation

Step 2: Creating A Solution Design

Often one of the most important yet overlooked aspects of digital analytics is documentation. Documentation provides an organization the ability to clearly define and record key components of its digital analytics implementation. At Analytics Demystified, we recommend starting the documentation using Excel as the format and expanding with additional worksheets as the requirements, Solution Design, and other components (e.g., QA processes) evolve.

Companies can rely on internal resources to generate documentation, or if using an agency or consulting partner, ask them to provide documentation  that should serve as the foundation for your analytics implementation. At Analytics Demystified we typically generate a Solution Design as part of our engagements and require that employees on the Digital Analytics team intimately familiar with this document because it will serve to answer all questions about data availability from the analytics platform.

Example Solution Design Documentation

Step 3: Developing Code

Unlike traditional development, digital analytics (especially Adobe Analytics) requires its own specific code base that includes events, eVars, and sProps to work properly. Most often we see clients outsourcing the development of code to external consultants who are experts in these specific technologies as this technical component of the job is often lacking within an organization’s core competency. However, in the long-term employing a Technical Digital Analyst with experience developing code for SiteCatalyst would position the company for self sufficiency.

Also, in the event that Tag Management Solutions are employed, a Data Layer is required to make appropriate information available to the digital analytics solution, which should also be addressed during the coding stage.

Step 4: QA

As with all development projects, digital analytics requires QA testing to ensure that tags are implemented correctly and that data appears within the interface as expected. At Analytics Demystified, we have developed our own processes for administering QA on digital analytics tags. Because QA requires input from technical analysts and IT developers, the process is typically managed via shared documentation (we use Google Docs) that can be accessed and modified by multiple parties.

Beginning with a QA Overview, companies should identify QA environments and Build environments with associated details on the platform (e.g., desktop, mobile, etc) as well as the number of variables to be tested. It is also helpful to develop a QA schedule to ensure that all testing is completed within development cycles and that both Technical Analysts and IT Developers are aware of the timelines for QA testing. Additionally, using a ticketing system will help Technical Analysts to manage what needs to be addressed and where issues are encountered during the QA process. The very nature of QA requires back-and-forth between parties and managing these interactions using a shared spreadsheet enables all parties to remain in synch and for work to get assigned and accomplished as planned.

Step 5: User Acceptance & Launch

Once the code has been QA’ed by the technical analytics team, it moves through the process workflow back to the business user who requested the tagging for final approval. While this part of the process should be managed by the Analytics Technical staff, it’s incumbent upon the business user to sign off on the tagging such that the data they will receive will help them not only measure the digital asset, but also make decisions on how to improve and optimize the asset.

A best practice at this stage would be for the digital analytics team to provide example reports so that the business user knows exactly what data they will receive and in what format. However, due to time constraints with development projects this isn’t always possible. In these cases, simply showcasing the prioritized requirements and the expected output should be sufficient to showcase what the data will look like in the production environment.

In closing, there are many different processes that can (and should) be applied to digital analytics. By building process around mission critical tasks, businesses can create efficiency in the way they work and bring new levels of standards and accountability to staff. By creating a process for new analytics requests, we’ve witnessed that organizations become more skilled at deploying tagging and reports in a timely manner with fewer defects.

Now it’s your turn…do you use a process for analytics? I’d love to hear how yours works.