A tracking plan is a document, typically a spreadsheet, that links your data & analytics strategy to the day-to-day collection of data. The tracking plan, however, isn’t a one-time project but a living document that should be revisited whenever there’s a change in the strategy. In a nutshell, it aims to clarify three things:
The best marketing is the one that feels 1:1. However, achieving this level of personalisation requires learning extensively about your customers. Plus, navigating this space can be a whirlwind given there are entire industries offering a range of different tools. Setting up a tracking plan will help you organise your efforts by having a single truth document to provide you with a much-needed overview. Whether you're only working with Google Analytics or using multiple tools, having a tracking plan will provide a reference for everyone in your team about what you're doing today and where you can take your business tomorrow.
The tracking plan can also work as a powerful stage-gate to ensure that you only implement tracking for those events that have a real value to the business. By formalising the question; “what value will tracking this event bring to the business?”, you will avoid implementing code that won't be utilised. This will focus your engineering efforts and reduce any waste by limiting scope creep.
The tracking plan also connects people across your organisation. As a centralised document that cuts across multiple functions, it requires collaboration. It provides a platform to have open conversations between management, engineering and marketing that is all rooted in what value it will bring to the business. Besides supporting better insight-gathering, this will also help break down skepticism and drive the organisation towards becoming more data-driven. To truly become a growth organisation it’s important to develop and nurture these touchpoints that connect people.
So now you know what a tracking plan is and why you need one. But how do you go about actually creating one?
Before setting up the actual plan, you need to have formulated your top business objectives. This will later help you answer why you should be tracking each event. You need to ask the key questions that provide you with the insight to skyrocket your business because ultimately, it isn’t about what you can measure but what you should measure.
As a cover sheet to the tracking plan, you should add your data & analytics objectives. This will help you keep what you’re trying to achieve top of mind as you complete the tracking plan.
When setting up your tracking plan there are several tools to use. Excel / Sheets is the quick and easy to use alternative but will require some extra discipline to keep up to date & clean. Airtable, which will require some extra work to fill in, provides you with the benefit of using relationships between elements (e.g. ensuring consistent naming for properties). As for 3rd party tools we mainly recommend using either iterative.ly or Trackplan. Both are purpose-built services for tracking allowing for maximum sense-checking & validation. If you’re using Segment, you can use the Protocols add-on for your tracking plan which immensely helps with QA and data quality as it alerts you and can even block non-conforming data.
In the end, no matter what system you choose, make sure you maintain the tracking plan - nothing is worse than an out of date tracking plan.
Best practice for setting up your tracking plan is separating it by purpose, for example a purpose may be to learn who your customers are or what they’re doing. For each purpose, there is a 'call' type that will help you answer these questions. There are two essential 'calls' (tracking and identification) to include. The identification 'calls' are used to answer, “who is the customer?”, while the tracking calls answer “what they are doing?”. However, there are some other frequently used 'call' types you may extend your tracking plan with that we will go over towards the end.
Tracking calls record the actions that your users take and capture relevant properties to describe those events. Some examples of events are; user signed up, invite sent, supplier viewed and product clicked. A key principle to have in mind when designing your events is to limit the number of events as much as possible and instead to enrich them with properties.
We recommend splitting up the event tracking into client and server-side events. A client is whatever the user is using (web browser or mobile) to browse the internet while the server is the computer that is providing the page that the user is viewing. Server-side events aren’t affected by adblockers but can only track identified users. These are for those critical events you need for reporting or email flows (especially STOPPING email flows). Client-side events can track anonymous users but are affected by ad blockers. Most (re)marketing platforms and conversions tracking systems rely on client-side events as they do their own user-identification. Even if they offer a server-side API (like Facebook) for conversion tracking you will capture a smaller percentage of your visitors. So since what and how you track your customers will differ if it's client or server-side, it’s useful to split these up in the tracking plan.
The document starts by covering the different channels (e.g. B2B, B2C) and why each event should be tracked to recognise what business value it holds. This then translates into covering what user actions will trigger the event and the properties and associated value that are used to enrich the information. The code snippet that is to be implemented is then added to ensure that that purpose is properly translated into actual implementation.
If you have events that are transformed before being shown in different destinations, such as Google Analytics, we recommend adding columns that show how these events will appear in different destinations.
Lastly, there are three columns for ‘Ready’, ‘Installed’, and ‘Tested’. Adding these columns will provide everyone in the team with a status update on the different implementation tasks that cut across functions. In addition, it works as a stage-gate that everything is developed and implemented as intended before moving ahead to minimise any issues that may arise.
The identification calls are used to associate the actions (or events) on your website to your customers. The basis for these calls is a unique user id and then different properties, or traits, of these users.
The template plan for user identification follows the same format as for the events. The main value of splitting the two is that they serve two distinct purposes; the event tracking focuses on actions while the identification focuses on traits.
Besides calls for identification and tracking of events, there are some other top-level calls that may be of use.
Page calls are used to capture what page your customers are on while screen calls are for identifying what application they are using to view it (e.g. mobile or web). Group calls capture a larger belonging of your customers such as accounts or organisations. Alias calls are for recognising if the user has any past identities as not all systems support changing a user-id once defined. To avoid this problem entirely we generally recommend having immutable user ids and not using for example email.
Setting up a tracking plan may seem a daunting task but it is an effort that will greatly pay off. Here are some additional best practices to keep in mind to help set you off on the right foot from the start.
Setting up your tracking plan will take some work, but it forces you to think about your data and ask important questions about your business and customers that will greatly pay off in the short and long term. Whether you’re working with a CDP like segment or GTM, a tracking plan will provide you with the structure to get the full benefit of your data.