The scene is all too familiar: Another marketing data request comes in through Jira. After a few back and forth conversations with the submitter, you eventually align on the exact data they need and are able to write the SQL query to pull it all together.
You’re nearly ready to hit send and move on to the next item on your list, when an update to the ticket arrives…actually, now that they think about it, marketing needs to change their original request.
In fact, not only do they need their original list of loyalty program members with a minimum spend of $300 in a particular product category, who rank in the top 5% in terms of customer lifetime value – but they also need to know, how many (if any) of those customers have a high email open rate in the last three months AND have not visited the website in the past 30 days. And if that audience is large enough, they’re going to want that list as well.
But do you even have that data? Can you be sure that data is up to date from all possible channels? What emails did marketing send during that three month time frame? Time to go back to marketing…
Does this headache sound familiar? For many of our clients, it was their reality for a long time.
Too often there’s a gap between what marketing needs and what data teams can readily provide in a timely way. And it’s easy to pinpoint the villain: most marketing tech and data tools don’t make it easy for non-technical marketers and technical data teams to collaborate, thus resulting in these infamous data silos and bottlenecks we all know too well.
But the stakes are high. Markets—and customers—move fast and a days-long delivery time for a request like the one outlined above could be the difference between making or missing a window of opportunity for additional revenue.
Unpacking the marketing-data request process
The data request process is nuanced – and its complexity depends on your business size, structure, the people and teams involved, and your data and marketing infrastructure. That said, after working with hundreds of consumer brands over the last decade, we feel confident that we can accurately summarize the process into four distinct stages:
Stage 1: Defining requirements
The data request process begins when marketing identifies a segment they want to target, and then identifies which data, attributes, events and conditions define that segment. This includes specifying the type of data, timeframe, granularity, and any specific filters. Accurately articulating these requirements is critical to ensure the data team understands their exact needs.
Stage 2: Submitting the request
Once the segment is defined, marketing submits their request to the data engineering team. This step involves documenting and submitting the request through a formal channel, such as a project management tool like Jira, or a dedicated data request system.
Stage 3: Assessing the request
Upon receiving the request, the data team evaluates its feasibility and complexity. They assess the availability of the requested data, potential technical challenges, and the required effort for extraction, transformation, and delivery. The evaluation stage helps determine the time and resources required to fulfill the request.
Stage 4: Delivery
After the assessment, the data team proceeds with extraction, transformation, and delivery. They retrieve the requested data either directly from their cloud data warehouse or from various sources, apply necessary transformations or calculations, and deliver the finalized dataset to the marketing team in their desired format. Voila!
Understanding where things go wrong
So, let’s be honest: the data request process almost never runs that smoothly and far too many data engineers thus find themselves mired down in writing ad hoc SQL queries, managing APIs and running ETL processes for what seems like an endless string of marketing requests. Consequently, on the other side, marketers feel like they’re being held hostage by the time between request and ultimate delivery of the data.
There are, of course, a few common culprits for these workflow issues.
Bottleneck 1: Lack of Clarity
One of the biggest blockers in the data request process arises when marketers are unable to clearly define their requirements. The resulting ambiguous or incomplete requests lead to misunderstandings, delays, and/or multiple rounds of revisions that will have all parties seeking out the nearest wall to bang their head against.
To overcome this, we have a few suggestions. First, an easily accessible and up-to-date data catalog or data dictionary enables marketing users to browse and discover what customer attributes and fields are readily available to them, and eliminate the incidence of repetitive data requests.
Data dictionaries have been around for a long time and exist in many formats. Some organizations build and maintain them in an internal wiki. Some are content with a simple spreadsheet. Most of our clients find value in using the data dictionary that’s built directly within our CDP, as it prevents them from having to reference multiple tools to build segments. Regardless of how the dictionary is built, it should be a resource that is owned and maintained equally by both marketing and data teams with each entry including the definition, key fields, use cases, and acceptable values.
Aside from referencing a dictionary, marketers should also invest time to precisely articulate their needs by providing examples or use cases when submitting their requests. This is one area where, yes, collaborative discussions with your team and marketing will help ensure mutual understanding and alignment. After all, you’re a data engineer not an order-taker.
Bottleneck 2: Limited resources and competing priorities
Another reality check: Data teams are in high demand and your work supports many departments across the entire company, not just marketing. If limited resources and competing priorities are getting in the way of fulfilling marketing requests, this is where the presence of savvy leaders on both marketing and data teams is key.
These leaders do the work of ensuring both groups are aligned with the greater business priorities. They can also introduce joint initiatives and practices such as regular alignment meetings (even if just quarterly) to strengthen the relationship between the two teams. Marketing leadership can also help instill the practice of providing clear justifications for the urgency of their team’s data requests.
Opening the lines of communication between teams will also help with negotiating timelines and managing expectations for individual data requests. A shared Slack channel, standing check-ins and even incorporating a 15-minute intake meeting after receiving a request are all meaningful touch points to consider introducing.
Bottleneck 3: Technical challenges and complex data and marketing ecosystems
Even in today’s business environment, where data is king, we still see businesses operating with their marketing and customer data scattered across multiple systems, databases, and platforms. In addition to some obvious governance and security issues, this also introduces all sorts of technical challenges to data engineers tasked with integrating and transforming data from diverse marketing sources.
The onus here is on marketing, who must work closely with your team to understand the underlying data architecture, provide assistance in data preparation, and commit to bringing on marketing tools and technology that fit into the larger picture of your data infrastructure.
Connected Segmentation: a better marketing request process
Ultimately, we’ve found that the best way to streamline the process of getting data to marketing is to eliminate the data request process completely.
Our brand new Connected Segmentation feature leverages a native integration between our CDP and Snowflake’s cloud data warehouse (CDW). With this new feature, Simon enables non-technical marketers to build customer segments directly within Snowflake, all without writing a single line of SQL.
With Connected Segmentation, the results of a given query reside right in the CDW. This means if marketers need to make changes to a segment, they are able to do so without needing any additional support from your team. In addition to this self-service quality, Connected Segmentation offers additional benefits to both data and marketing teams, including:
- Better security and governance: Connected Segmentation solves the issues introduced by decentralized marketing data by bringing the task of building customer segments into the data warehouse, rather than in separate marketing tools and workflows.
- Better analytics integrations: Because Simon now integrates with all the other tools built on the CDW, such as business intelligence tools like Looker or Tableau, these technologies can now work together to drive incredible insights, and enable marketers to change course quickly if there are performance issues.
- A new (and actual) single source of truth for all business data: Simon’s fully Snowflake-Connected CDP establishes a true, single source of truth for your organization: your cloud data warehouse.
- Access to new marketing data products: Connected Segmentation creates a direct feed of marketing campaign data–events and engagement data, and performance analytics–right back into the data warehouse. This open back and forth flow of data enables marketing to contribute to your data mesh, and leverage additional value from the data their work produces.
In practice: a Snowflake + Simon CDP success story
Prior to Simon, Travel+Leisure Co, the world’s leading membership and leisure travel company, was plagued by heavily siloed data. To overcome this challenge, the company used ETL processes to bring data into an on-premises database and managed data in batches. While that worked for a while, the process required a ton of maintenance, and importing data in batches prevented the team from being able to serve experiences in real time.
Enter Simon and Snowflake. Today, Travel+Leisure centralizes its data in Snowflake and leverages Simon’s marketer-friendly UI to deploy data models that prioritize messaging based on customer intent, and unify their 1:1 personalization efforts across all channels. Travel+Leisure’s marketing team now enjoys a significantly more simplified process for building segments–which now update automatically and in real-time. They can aslo seamlessly deliver cross-channel campaigns to those customer groups – across their website, digital ads, email, call center, and direct mail channels. Finally, the company now can leverage a continuous flow of data back and forth between its marketing channels and the CDW.
All of those points speak to a tighter integration between data and marketing, but it’s also important to know that within a year of implementing Simon and Snowflake, Travel+Leisure reported over $350k in incremental revenue.
It’s time to go from order takers to strategic partners
It’s true that with a better data request process, marketing and data teams can start their conversations from the same page and at a higher point of departure and they can have higher value interactions starting from a shared understanding. But do you even need the process after all?
Having a CDP sitting between marketing and the data warehouse not only alleviates bottlenecks in the flow of data to and from marketing, but opens doors for data and marketing teams to have better conversations. And, after all, isn’t the goal to partner with marketing to solve complex business problems?
Connected Segmentation is a time creator – enabling more time and resources for your team to do the work you love: building models, uncovering insights, asking more and better questions, and bringing ideas to marketing.