Welcome to the next installment of the mythbusting series brought to you by INCRMNTAL. In this edition, Arun Srinivasan, CEO/co-founder of Clarisights, dives into the topic: With scales comes structure.

In this article, we want to debunk the prevalent myths and misconceptions around whether the most scaled up enterprises actually have the best structure - for data, for teams, etc. And, if not, why is that?

Given Clarisights specializes in providing better structure around data and associated marketing reporting for large enterprises (ranging from Uber to Universal Music Group), we felt it was natural to invite their CEO Arun to write on this topic.

In our mythbusting series, our mission is to unravel these types of misconceptions, one by one, and shed light on the truth. If you would like to check out the first piece on incrementality and user-level data you can check it out here.

The assumption is that, as businesses scale, they are able to build what they need internally and create better structure for these needs. We tend to assume that mature companies have better processes, better data structures, robust workflows, and advanced tools to support their business and teams.

This is true in many cases. Amazon first built out AWS as an internal system; Spotify first built Backstage as an internal-only developer tool, ahead of open sourcing it and now making it commercially available; Meta is well-known for building their own data centers.

There are many other examples from some of the most disruptive companies. Even so, with scale comes chaos, mess and unnecessary silos. A good example is what happens to data availability and reporting as organizations scale.

In the current environment, growth marketing teams are under increased pressure to show work and drive impact. Growth marketers have access to a virtually endless supply of data across marketing performance in various channels to support this. This data can be sliced and diced in many ways to generate insights to inform actions that lead to more traffic or conversions at lower costs. At scale is when the true challenges begin.

Neither the tools marketers use nor the team structures are built to take best advantage of this data. As a result, many opportunities go unnoticed. Teams suffer from inefficient investment and use of resources, especially since getting your job done often becomes more important than data democracy and cross-org info sharing. Without tooling positioned to overcome this, marketing teams and their investments suffer.

In this blog post, we’ll talk through why this occurs, the common challenges, and the impact this has on team’s and their strategies. We’ll then cover how some teams have approached addressing this.
 

The Enterprise complexity

Businesses become more complex as they evolve to scale up and enterprise stages. In addition to growing headcount, they add more business lines, sell more products, expand to more markets, etc.

For Growth Marketers, this all means more opportunities to drive positive results. More marketing channels, campaigns and creatives, etc. This often means billions, or even trillions, of rows of data owned and managed by various teams across the org.

In an environment in which budgets are under increased pressure from internal stakeholders, growth teams are expected to prove the incremental impact of marketing initiatives. This means, within all this data, marketers need to find pockets of efficiency to either reduce acquisition costs or increase sales - ideally both. They need to constantly test and find arbitrage opportunities.

Marketers empowered with data are at the core of accomplishing this.
 

The Modern Data Stack and its data silos

As businesses grow, they inevitably adopt some form of the Modern Data Stack (MDS), relying on a general purpose BI tool (such as Tableau or Looker) for their source of truth reporting. Organizations invest vast resources, both in tooling and people to stitch together tools and to build and maintain this “reporting stack”. The hope is that this will answer all marketers’ reporting and analytics needs and take advantage of the opportunities in the data.

The reality is that a standard BI tool is excellent for KPI snapshots and executive-level overviews; however, the stack of tools stitched together has a number of gaps that prevent capturing this opportunity. These gaps include:

• Visualizations with truly granular data over long timeframes are often slow to load since queries have to scan hundreds of millions of rows to return results; 
• It gets prohibitively expensive to allow for flexible and performant evaluation of granular data (creative asset-level; keyword-level; etc.) at scale, forcing data teams to pre-aggregate data;
• Analytics/BI teammates build new dimensions or metrics for marketing since this requires SQL (or coding proficiency) and complex data models that marketers can’t make changes to. This limits marketers from moving fast with data exploration, testing cycles, etc. on this reporting stack;
• Using dashboarding tools as a reporting UI makes it difficult to quickly adjust or make changes, even for something as simple as changing a time period bucket from “weeks” to “days”.

Dashboards are considered the key end-product for many data teams. They're great until marketers ask for a filter that is missing or a cut of data that they need. This then leads to what we often see referred to as ‘the spreadsheet problem’ in order to fill this gap in the modern data stack.

Spreadsheets are flexible and also familiar to every marketer. As a result, nearly every enterprise growth marketing team leverages spreadsheets extensively. While this can be an effective stopgap measure, it has some challenges, including:

• Spreadsheet data needs to be updated manually, at very least, daily. This is typically time consuming
• Every spreadsheet report is another data silo, making a single source of truth for insights nearly impossible;
• Formulae are prone to break when something changes in the data. Because they are updated manually, they are more prone to errors than a managed pipeline
• VLOOKUPs, SUMIFs and other formulae need to be adjusted every time a marketer wants to make a change to a dimension or metric, or to create new cuts of data.

This often leads to data sanity and accuracy issues. Spreadsheets are every data team’s governance nightmare.
 

The cost of siloing your marketing tactics

Challenges from scale are not limited to the tools. The data silos inevitably drive more siloed team structures, limiting collaboration. This makes it difficult to evaluate different tactics holistically, such as:

• How does my TV flight impact search traffic, site visitors and overall purchase conversions?
• What is the correlation between Paid Acq and CRM campaigns, or, from a KPI perspective, my CAC and LTV across channels & campaign types?
• Am I ensuring that my Paid Acq messaging feeds into email messaging and then into my landing pages? Are there gaps that I am missing?
• What creative concepts have provided the best ROAS in the last 2 years?

The failure to connect these dots means teams will operate with blinders, solely focused on hitting their own goals. They won’t consciously think about the impact of their day-to-day work to the broader business. Marketing teams will over-invest in certain areas and overlook others.

Imagine you are a paid acquisition manager for a food delivery business. You’re continuously scaling up Paid Acq campaigns because you see CAC reducing, while overlooking that these customers are not engaging with CRM campaigns 1 month later. The data infrastructure needs to facilitate monitoring this seamlessly, including layering other datasets (cohorted revenue, CLTV, etc.) for analyzing campaign performance holistically. Otherwise, you are likely to miss this insight and misuse marketing budget.

This can cost businesses millions, or even billions, in sales, and be the difference between winning and losing.
 

Team structures are evolving, but is the tooling?

Some of the best growth stories of the past decade have addressed some of these challenges with, what is known as, ‘Pod structures’. This includes Lyft, Pinterest, Uber, and Airbnb, and has helped teams drive cross-functional collaboration on key initiatives.

Sigal Bareket, Lyft’s former Head of Growth Marketing, shared the following with me:

“It's essential to have teams that are experts in different disciplines. However, problems start when these experts try to solve business challenges using only the tools their discipline offers and are not looking at the same data.”

While we see progress in team structures, this is largely not being addressed in the data and reporting space. Businesses are actively assessing how to make their marketing teams and investment more efficient in this tougher growth environment. In order to do this with even greater success, they’ll need to lean into finding solutions outside the traditional tech stack.