- Massive amounts of marketing data create a morass of challenges within a business that require experienced engineers, data scientists, and data visualization analysts to unpack
- Marketing technologies are starting to solve for the marketing data value solution but typically attack only one or two of the core challenges between engineering, analysis, and visualization
- Without solving for all three realities, marketing data will continue to underperform against the investment made in collecting it
Extracting value out of marketing data has always been a challenge. Typically, there are three core competencies that are needed to unfurl the latent potential of massive amounts of marketing data.
- Engineering to collect, extract, normalize, join, and process large data sets across a spectrum of channels, technologies, and locations.
- Data Scientists and Analysts that intimately know your business and can create meaningful queries, algorithms, and models that will automate the value generation process.
- A team of Data Visualization Analysts that can interpret and present the data in universally understandable ways (we’ll cover data obfuscation and privacy in another piece).
This array of competencies isn’t native in most organizations, either due to resource limitations, skill deficiencies, organizational alignment or any combination of these. Yet, most businesses understand and yearn for tools that can put them at the heart of the Reuleaux triangle of engineering, analysis, and visualization without handcuffing their operations and budgets.
Each day seems to bring forth a new set of tools that looks to resolve these challenges. There are many data visualization tools that are quite compelling. These tools have replaced the need to custom build complex business dashboards that were often too difficult to interpret or maintain, but these tools still require skill positions to develop meaningful interpretations. There are an increasing number of data management tools that aggregate, normalize, and join large data sets, but still have difficulty crossing borders like known and unknown. Finally, there are a growing number of tools that try to automate the interpretation of data through canned models or algorithms yet still fall short as a way to answer simple questions specific to your business.
What this all means for marketers is an ever growing set of complex realities that they will need to navigate if they hope to achieve any form of value from their marketing data. So let’s take this three legged stool one leg at a time.
Reality 1. Marketing data is increasingly siloed, fragmented, and immovable. When once we looked for data transportability we must now seek application transportability. More and more data sets will live in a state of rest within its respective environment where the core data can never leave, the analysis must occur within its walls, and the extract must be obfuscated. What is therefore required is a technology solution that can move to where the data is located, query within its walls, and pull cohort level analysis that can be utilized across environments.
Reality 2. Unlocking the truth behind the data is difficult under the best conditions and the more robust and complex the data sets become the more resources this will require. It has been apparent that for the last few years data science is the place to be. Given the number of schools building programs around it it’s clear that this is a long term career opportunity. This doesn’t solve the short-term need marketers have to solve real challenges in building the models and algorithms necessary to power value generation from their data. Without enough meaningful resources holding up this leg of the stool marketers will continue to achieve surface level value from their data and never truly unlock the uncommon commonalities hidden in their data sets.
Reality 3. Democratizing access to meaningful renderings specific to job function and role is critical. There are countless new and extremely powerful tools to furnish data and build compelling dashboards. For the past few years the problem has not been in the technology but how that technology is wielded. Templates and canned dashboards will typically only add to the confusion because they are trying to do too much for too many. Ultimately, marketers will require self customizing dashboards that answer specific business questions in ways that are meaningful to the role or function of the end user.
There is an old adage that says good, fast, cheap … you can only pick two. In this case, without effectively covering all three core components of the data value exchange: engineering, data science, and data visualization analysis, marketing data will continue to suffer from the cycle of over investment in limited solutions. In future posts we will discuss the nirvana fallacy around marketing data and ways you can begin to truly extract the value you need from your data assets.