How to audit a company’s marketing analytics infrastructure (+ a checklist to do it yourself)
How to audit a company’s marketing analytics infrastructure (+ a checklist to do it yourself)
We cover the different types of data assets that should be audited, when and how often you should do audit them and what you should have in hands at the end of your project.
Marketing analytics and marketing data are umbrella terms that can encompass many, many different activities. For this article, we’ll cover three specific use cases for marketing data:
Measurement: To quantify campaign performance with the goals of extrapolating ROI, understanding impact on indirect metrics (e.g. on the product funnel) and comparing campaigns as close to “apples and apples” as possible.
Insights: To derive insights from your own marketing campaigns that are valuable for other marketing and product teams; and can later on be used to drive business results. (For example, with paid media testing.)
Action: To leverage data to improve campaign performance by reaching the most valuable audience with effective messaging. For example, by using a RETL.
Together, these three use cases influence how marketers allocate budget, define their strategies and improve the performance of their (lifecycle, paid media, content) campaigns.
But how do you make sure your marketing teams have the right data assets to execute on their vision? That’s what we’ll aim to tackle in this guide.
What are the signs you should conduct a marketing analytics audit now?
Timing is everything. The best marketing analytics audit is timely, conducted when it can be most beneficial for an organisation. Scenarios that call for an audit are diverse:
Complaints about data inconsistencies or things constantly breaking in the system
Example: Your paid media campaigns keep needing to be relaunched (and back to “learning periods”) due to broken tracking.
Possible Cause: This can be due to a lack of marketing data engineering best practices, such as using a centralising tag management system, dbt, server-side tracking or alerts.
Discrepancies on how teams calculate metrics
Example: Different teams, like sales and marketing, may have separate reports that supposedly track the same metrics. However, they show different results, leading to debates about which report is correct.
Possible Cause: This often arises when different tools or platforms are used for similar purposes without a centralised approach or when there's a lack of clarity on why discrepancies exist.
A blockage in marketing initiatives due to a lack of insightful data
Example: Your content marketing team is ready to create a new quarter's worth of material. However, they're unsure about which topics resonate most with the audience because past data is either missing or inconclusive.
Possible Cause: This could stem from not having appropriate tracking for content engagement metrics or from not gathering feedback and preferences directly from the target audience.
Marketing keeps needing to compromise on their vision
Example: An onboarding campaign can’t be customised based on customer data, so instead all new users get the same messaging at the same time.
Possible Cause: A lack of appropriate data tooling that enables marketers to execute on their vision, such as a Customer Data Platform or a Reverse ETL.
At times, you might even find the need for an audit when you're already aware something isn't working. Addressing these preemptively can save countless hours and resources in the long run.
What you should have in hands after the audit?
A marketing analytics audit isn't just an exploration—it's a blueprint. Think of it as a diagnostic report that delves deep into the health of your data, uncovering strengths and spotlighting vulnerabilities. At the end of your audit, you should have:
Identified sources of data inaccuracy or incompleteness
What: A thorough examination of where and why your data might be falling short.
Example: The audit may find that your leads form doesn't capture all the required user information because certain fields are optional. This could be leading to incorrect lead qualification by your sales team.
Highlighted gaps in your data structure and data stack
What: An assessment of any oversights in your current tools, platforms, or way of working that might be causing inefficiencies or blind spots.
Example: The report might highlight that marketing people are unable to self-report on their campaigns and require either a dedicated dashboard or SQL-training.
Uncovered opportunities for more effective segmentation and targeting.
What: A section dedicated to potential improvements in your audience targeting, backed by the data insights from the audit.
Example: The report might highlight that adding a Reverse ETL to your stack might enable the lifecycle team to create better performing emails.
Translated your discoveries into an actionable roadmap
At the audit's conclusion, you should craft a roadmap document that's both diagnostic and prescriptive, ready to guide the next steps in refining your marketing analytics approach.
Who should you involve to execute your audit?
A comprehensive audit needs to be executed by two types of roles: a generalist marketer with knowledge in the different types of marketing activities (lifecycle, paid, etc) and a marketing analytics expert who understands platform-specific nuances.
Sometimes, the roles of the marketer and the marketing analytics expert can blend. I've found myself wearing both hats and this is precisely the sweet spot where I thrive: merging the creativity of marketing with the precision of analytics.
Regardless how you approach it, don’t overestimate the human impact on data. During the audit, you should speak to different marketing teams to understand how they use (and don’t use) data.
What are the questions you should be asking when performing a marketing analytics audit?
The following data assets and data structures should be audited to assess the state of your marketing analytics:
Audiences, segmentation and customer data
What are the different audiences or audience segments that are used by your sales, lifecycle and paid media campaigns? How complete and accurate are these audiences? How often are they updated, cleaned or refreshed? Is there additional customer data (macros, custom variables) that can be passed to enrich audience segmentation or customer messaging?
Tracking events and corresponding variables
What are the different steps taken by new users until they convert? On what tools are these events tracked? Are the events tracked consistently across tools? How are these events used to measure or enrich campaigns? Are there additional events that could offer additional benefits? Are there variables that could be useful to be passed alongside the event?
Campaign and UTM naming conventions
Are campaigns named consistently within the platform? Are UTMs named consistently across different platforms? What are the different fields that are included in the naming convention? Is the naming convention harming the campaign set-up or optimisation in any way? Are there additional naming convention fields that could be useful to derive insights from the campaign? Are we able to join disparate data sets (e.g. ad server, DSP, Snowplow) with the naming convention?
Data Access
Are stakeholders able to access the data they need to make decisions? What’s the process for accessing this data? Is the process from them to extract and analyse data leading to human errors? Are they able to derive insights from the data? Do they trust the data? Do they find the data complete? Where do they think campaigns can be improved with data? What are the hypotheses they can’t validate and why?
Time to get started on your audit with free checklist
Marketing data should enable you to run the campaigns you desire, understand how they perform and derive insights for business decisions. Conducting a marketing analytics audit will help you understand where you are today, identify what you need to tackle and map how you will get there.
I’ve created a Google Sheet checklist with the steps to take for a successful audit. All activities are organised within a period of 11 weeks, so you can start and finish your audit in under a quarter.
Just request access and get started:
