The Specifics of Assessing the MarTech Stack
By Josh Kaplan
No matter the size or financial health, many companies continually stunt the growth of their Martech and Data Science stack by exacerbating existing stagnation. Often issues or roadblocks found in one technology tend to be looked at laterally, with the thinking that a newer, better version of the same product will solve everything. To marketing SME’s, this is the equivalent of trying to remove snow with more snow.
With the emergence of Silicon Valley startups and niche technology companies, there are more options than ever to solving a specific problem. Coupled marketing technology suites are going by the wayside, making room for specialized tool functionalities and narrower focus. The ability to diagnose the business challenge, analyze the available data and tools, and fill in the blanks has become more of a puzzle than a straight line. For this reason, proper technology assessments have been paramount in navigating these increasingly muddy waters.
Using a sandwich analogy, when assessing the Martech and Data Science stack of a given company it is best to start with the buns and let the technology fill in the meat and vegetables. In other words, start with what is available to you at inception, coupled with the knowledge of where you want to end up, and use data and technology to bridge those gaps. To do this, the first step would come in the form of a discovery phase.
During multiple discovery sessions, analysts have the ability to dig deep into the business problem, assessing ways similar problems were addressed in the past, analyzing the technology currently used to achieve goals, and diagraming the workflow in which the marketer used to get from point A to point Z. Additional questions will arise about available data, namely where it is, how quick and easily accessible it is, and data health such as formatting and cleanliness. Questions around technology, the marketer’s thoughts on it, needed features, and performance analysis help the analyst understand just what is needed to solve the business problem efficiently and accurately. Finally, an assessment of downstream applications of the data will help to cover all bases as it pertains to what the marketer wants to do with their data.
Next, a deeper dive into the technology used is of necessary importance. Learning which systems are being used to accomplish key tasks and goals, and understanding their strengths and weaknesses, are key steps in this process. Generally, technology questionnaires accomplish much of the heavy lifting, while learning and training sessions with various users help the analyst to understand how things are being used and why. This applies to all phases of the project, from data prep tools, data science stacks, and marketing automation tools to DMPs, CDP’s, campaign management tools and the like. The result of this is a clear plan for how to go from a finished data set to a robust marketing program.
Finally, a data audit will really help the analyst understand what data is available to them. In all Martech and Data Science projects, the most important piece is the data itself. The context gleaned from understanding initially what it looks like in its current state, how much preparation work is needed, and understanding which data points could be filled gives necessary insight into the data preparation process, resulting in an air tight scope.
A good consultant will have a strong opinion on technology but will remain flexible enough to adapt to client legacy systems. Here is an example – A fortune 500 company is looking to create a golden record data set from disparate data sources and use that set to serve programmatic display ads downstream. Suppose you go to that company and uncover that they have a DMP and an address standardization tool internally, plus a lot of data in excel spreadsheets and various DB’s. An analyst will know to solve this business problem, you will also need DSP technology, deduping and fuzzy matching technology, and a CDP or CDI solution to make sense of all the various data. By assessing the technology, you could potentially uncover that the DMP does not have a robust DSP connection, necessitating one for the client. Your discovery phase would allow you to use business rules and context to know which data preparation tools may be needed. Your data audit phase will help guide you how to fill in the gaps to complete the client’s data, giving you a fully filled out dataset. By going through these necessary assessments, you can complete the stack for a client, help them understand which tools are needed and which can be sunset, and devise a streamlined approach for solving the business problem.
