Fix Don’t Stitch: CMOs, Beware of the Data Frankenstein Monster You’re Creating

Fix Don't Stitch: CMOs, Beware of the Data Frankenstein Monster You're Creating

Impact LogoThe internet was promised to be a place of sheer marketing nirvana. And with the advent of digital advertising, brands were given the power of large-scale addressability for the very first time. Finally, advertisers could target individual consumers, with software capable of telling them exactly when and where each customer encountered their ad message. Over time, we built software systems that empowered us to use data to find our ideal customers and measure how well we reached them.

But today, most marketing stacks look less like this utopian vision of the future and more like Frankenstein’s monster: A mish-mash of partly-functioning tools that have been stitched together. Or as I call them: Franken-stacks. With so many parts in play, it can be difficult for marketers to get to the bottom of where their data actually comes from, or to put this information to good use. It’s no wonder only 26% of organizations say they have a solid data-driven marketing strategy, and only 29% of advertisers consider their measurement data completely trustworthy.

Read More: Keeping A Positive Approach to Data Management is Critical to Revenue Traction

Simply put, this isn’t good enough. With $1 trillion of spend each year, marketers need to come up with better solutions for simplifying the data collection, processing and reporting process. With the right strategy and changes, we can turn our Franken-stacks into Super-stacks.

Origin Story: How Franken-Stacks Came Into Existence

Before marketers can go about fixing their data stacks, it’s important to understand how things came to be like this.

Most brands begin their digital marketing journeys using multiple channels, such as paid search, email, affiliate and Facebook advertising. With each channel typically operating in isolation, companies start running into problems when CMOs begin asking the hard questions about holistic ROI. In order to answer them, marketing departments bring in all kinds of new tools: spreadsheets, proprietary data warehouses, attribution vendors and visualization tools, to name just a few.

This is the start of building a Franken-stack: A jumble of mismatched tools operating across multiple siloed channels, without a holistic view of the customer and lots of reporting errors. With so many systems in play, most marketers are unable to determine where their data comes from or who compiled it. This means they don’t have normalized event-level data that would allow them to access metadata regarding individual impressions, clicks and conversions, as events are not linked to their source. And worse, without proper data governance in place, the result is missing or inaccurate data, typically due to tagging errors and lack of organizational process. It is not uncommon, for example, to find conflicting vendors assigned and attributed to the same click tag!

What marketers are left with is aggregate, visualized data that looks visually appealing, but where actional performance metrics don’t exist, are non-transparent and often incorrect. A great looking visualization tool on top of bad data is like putting lipstick on Frankenstein’s monster — it’s still a monster!

Read More: A Publisher’s Journey to Data Collection: Adapting Netflix’s Approach

And with everything siloed across different vendors, tools and channels, marketers are also forced to make sense of contradictory data. For instance, what Google might count as multiple raw clicks, various vendors might count as a single click as defined by their methodologies for a session. Discrepancies are certainly the nature of the beast; however, marketers need to understand them before embracing and working to minimize their effect.

For all its fancy parts, these Franken-stacks offer no real clarity into marketing performance and no way for advertisers to confidently use their data to make smarter marketing decisions.

Building a Lean, Mean Actionable Data Machine: A Super-Stack

Fortunately, all hope is not lost. By rebuilding their infrastructure properly, marketers can develop a super-stack: A marketing stack capable of providing full transparency, holistic data analysis and powerful, actionable insights.

The super-stack starts with a big data foundation, a scalable location to capture data and make it accessible. This data lake will need to track and ingest marketing data at the event level (i.e., impressions, clicks and conversions), across paid search, Facebook ads, direct display, email and any other marketing channel, including offline. super-stack also need to have the intelligence and processing power to verify and normalize this data, including the deduping of events, users and other items. It is critical that marketers can compare apples-to-apples across their various channels of marketing spend. This normalization process also includes cleansing data of fraud, which now represents over 30% of all ads, creating more than $19 billion per year in waste.

Read More: What Data Orchestration Solutions Mean for B2B and B2C Marketers

Once the data is ingested, it needs to be accessible. And a super-stack supports the capability of algorithmic data modeling, which helps marketers gauge incremental business results. To this end, analysis and visualization needs to be both portable and interoperable with other systems. And in a world where the average North American consumer will have 13 connected devices within the next three years, it’s imperative that the super-stack can dedupe at the individual level, offering a person-centric view of each individual customer.

Once all these things are in place, marketers can put their super-stack to work by integrating an application layer capable of predicting the best way to allocate budgets across channels and placements. These algorithms can even figure out what to bid or pay at the event level, with a high degree of confidence. It all comes back to building the foundation of a super-stack with a clean, normalized data set — the holy grail for anyone attempting data-driven marketing.

Taking the Next Step

While untangling your Franken-stack may seem like a daunting task, building a super-stack from the ground up is the only way to unleash the true power of addressability, and achieve the elusive state of sheer confidence in making marketing decisions.

With the right data ingestion and normalization technology sitting at the core of your super-stack, combined with rigorous organizational data governance, you’ll be well on your way to an integrated system that can understand marketing results at the event level and intelligently optimize future campaigns.

Remember that Frankenstein’s monster ultimately turned on its creator and they both paid the price. Fortunately, there are better ways to breathing life into our companies, that is, growth through better marketing!

Read More: Salestech Predictions 2019: Bigtincan CEO on Sales Enablement and Salesforce Technologies

Previous ArticleNext Article

Leave a Reply

Your email address will not be published. Required fields are marked *