The State of Digital Marketing in 2023

Improving the efficacy of marketing performance is a priority of every company in 2023. Marketing budgets are at risk, hiring is frozen, teams have shrunk, and procurement is pushing for tooling consolidation.

How can marketing do more with less? How can marketers continue driving incremental results in this economy and with its constraints?

The answer is, of course, with data.

In this post, I’ll examine the often-overlooked challenge of solving the last mile of marketing analytics: The business users adopting your dashboards and making use of your modeled data.

Data-Driven Marketing

Measuring the impact of marketing budgets in a constantly evolving landscape of channels, metrics, diminishing tracking ability, and performance levers is increasingly difficult. It has been, it is, and it will be complex.

The simplified two-step process to harness this complexity and drive results from marketing investments is as follows:

1. Building: Data Collection, Transformation, Analysis, Data Modelling, and Visualization

2. Delivery: Effectively delivering the modeled data to the business user

Building: Data Collection, Transformation, Analysis, Data Modeling, and Visualization

Modern data teams conduct incrementality testing and create sophisticated Marketing Mix Modeling (MMM) and Multi-Touch Attribution (MTA) models to guide the decision-making of marketing teams based on extensive data collection, analysis, hypotheses, and testing. Without the continuous testing and the data models, it would be impossible for marketing teams to create a holistic picture of the marketing performance and make educated decisions based on data.

To achieve this, data teams typically craft Looker or Tableau dashboards for the marketing teams by compiling data from various marketing channels, measurement solutions, internal data (budgets, targets, orders), and modeled data (attributed KPIs, MMM model recommendations). These dashboards are intended to be the single source of truth for the business user to make budgeting and optimization decisions.

Not simple, but relatively straightforward, right? This is what many data teams have already built.

But collecting the data, building a data model, and visualizing it in a BI dashboard is only the beginning. The challenge lies not in building the models but in effectively delivering them to marketing teams. Data collection, analysis, and visualization are arguably the "easy" and more exciting parts. The moment actual value is created is when users adopt the dashboards and make decisions on the models provided. However, due to some inherent limitations of the BI dashboards, this goal often falls flat.

The Delivery Challenge: Limitations of Generic BI Tools

BI dashboards are great for answering well-defined questions. However, for exploratory analyses or deep dives, they fall short. Data teams usually solve this using SQL, Hex, or notebooks alongside their dashboard solutions.

For marketers, the challenge is identical because digital marketing is not static. Numerous evolving variables and testable hypotheses exist that could either create a compounding long-term effect or an immediate significant impact on marketing performance.

Marketing budgets deliver incremental performance when marketing teams are empowered to make fast and accurate decisions. To achieve this, marketers combine their knowledge of the business, target audiences, all the peculiarities of their channels and markets, PLUS the data models built by the data teams.

To make the most of the budgets they own, marketers need fast, granular access to create insights, the ability to quickly test hypotheses, and the ability to make changes to the marketing reporting business logic. However, the generic BI tools have some inherent limitations inhibiting this process.

Because of their primary use case as business dashboards, Looker and Tableau are slow to load when drilling into more granular, creative, or keyword-level views, and changes are usually slow and manual.

But marketers require flexibility, and flexibility requires being able to quickly amend the logic by, for example, creating new dimensions, metrics, and visualizations. That's when things turn complex.

Making changes in a generic data stack may require changes in the data connector, data warehouse, and the dashboard tool. In the worst case, each change is applied by a separate team.

Consequently, the insights of these BI dashboards are aggregated to the campaign level, which is not detailed enough for modern marketing teams.

This is why data teams often face challenges with the marketing dashboard adoption and mobilizing the MMM and MTA models for everyday marketing decision-making.

This leaves marketers with two choices:

1. Raise tickets to the data team to perform ad hoc analyses or implement changes to the current dashboards, or

2. Build their reports in spreadsheets or Looker Studio (becoming "data janitors").

As the first option is often not feasible or fast enough, marketing users revert to alternative solutions like spreadsheets, standalone Looker Studio reports, or native ad platform reporting, where decisions are usually made in silos without the guidance of the data models.

"You can lead a horse to water..."

The Cost of Unfit Tools

Needless to say, it is incredibly costly for companies to first spend years of person-hours in building a reporting infrastructure with the generic BI tools and creating data models, just to learn that marketing won't adopt them.

Data teams spend significant time, money, and opportunity to do their best to deliver what marketing requires.

Marketing teams' marketing budgets are wasted because they don't have the right tools and insights.

And both teams often fight each other.

There is a Better Way

So, how to overcome these challenges?

The ideal solution…

✔️ Must have all the data from marketing platforms, measurement tools, internal data, and the modeled data
✔️ Must support strong data governance
✔️ Must be compatible with the Modern Data Stack
✔️ Must be built for the marketing user and support granular self-serve report building vs. being dependent on data teammates for any changes
✔️ Can't require heavy ongoing maintenance
✔️ Can't be overly expensive to run (in terms of HR cost allocation, data warehouse processing costs, or tooling) and must be demonstrably ROI-positive

Want to find out how some of the world’s most sophisticated data and marketing teams at Uber, HelloFresh, and Delivery Hero overcome this delivery challenge? Schedule time with us.