The important role data plays in the success or failure of agencies and marketers cannot be understated and yet, the role of consistent data taxonomy and implementation and use of a data dictionary, is underutilized. By codifying the streams of campaign data that flow into and out of an agency through a ‘data dictionary,’ marketers can begin to create scalable processes for their client reporting and better communicate the value and worth of their efforts.
We sat down with NinjaCat’s Manager of Learning and Enablement, Dustin Blackwell, and Enablement Specialist, Jen Musto, to hear about their experience with taxonomy, data dictionaries, and how both agencies and clients can better prepare themselves, and their data, for reporting.
The State of Data Taxonomy In Agencies
Since there are so many overlapping platforms and systems that comprise a martech data stack, there is a ton of data to sift through and even more heavy lifting and translation efforts are required to parse through the analytics and find what’s reportable and what isn’t.
“Most agencies don’t really have an overarching narrative when it comes to their campaign data,” says Dustin. “This results in a morass of data and spreadsheets and custom calculations that can be difficult to wrangle into reporting software. My advice to agencies is to know what kind of story you are looking to tell with your data, and then start to build your reporting templates using that narrative. The starter marketing report templates we have in NinjaCat provide a nice place to launch from.”
In addition to the amount of data coming off platforms and apps, there is the issue of taxonomy and naming conventions; what counts as a conversion event in Google Analytics might not line up with the definitions for conversion events on social media, or the client CRM, or any of the other places marketing data resides.
And then, as if there isn’t enough confusion around data taxonomy, client defined success metrics can often be misaligned with internal agency metrics as well as platform analytics, making the entire act of getting your reporting efforts and client dashboards on the same page a herculean effort.
“Client-defined conversions are more important than agency definitions, because these things can be so unique to business models,” says Jen. “A key part is understanding how clients actually make money. It’s easy to get lost in the niches - when you have to explain metrics, like page views vs clicks, and what those mean to the bottom line - this can be frustrating.”
Due to the misalignment in data collection that exists on all sides of the reporting process, what could be a repeatable and scalable activity for your agency and clients, can easily become a bespoke, handcrafted effort that requires a ton of one-off mechanisms and machinations. This leaves your agency in the reporting lurch as far as man-hours are concerned, and would no doubt leave your clients waiting while your team sorts through table after table, wrangling data, copying and pasting performance analytics into reports that may take hours or days to create, when they could take minutes.
“Proving the value of your agency’s services is essential to reporting,” explains Jen. “The real proof of value is guiding clients to understanding the analytics and actions and then advising them on next steps.”
The Benefits and Importance of A Data Dictionary
Since there are so many ways actions and engagements with advertising can be tracked online, it only makes sense that an agency would want to define conversion events in a holistic manner across all clients. But the reality isn’t so simple, says Dustin.
“If your end goal is tracking conversions, the thing that’s most important to include in your reports is conversion events, not the platform metrics from Google Ads or FB ads. Most agencies understand each client has specific metrics, this is why things like global metrics and global filters in NinjaCat can help switch conversion metrics across all clients making reporting scalable, rather than getting bogged down with bespoke dashboards for every client. It’s all about exceptions and unique customizations - are you making an exception for every report you create, or are you approaching the reporting problem in a one-to-one fashion?”
Another way a data dictionary can help agencies and their client reports, is that a consistent approach to analytics can even out workloads. Most agencies have a few big clients that tend to take up more resources, but this can skew efforts and abilities within the agency to deliver homogenous performance reports across their entire book of business.
“By just focusing on whales, you miss the chance to scale your reporting efforts to existing and potential clients,” explains Dustin. “The dream is that you build one template that powers the reports and dashboards you create for every client. A data dictionary makes this dream closer to reality.”
Jen had similar thoughts in regards to master templates and cohesive metrics for clients.
“You want to create templates that you can customize, not customize every template. The thing most agencies struggle with is “you have all these metrics, but what do they mean?” The agency has to educate the client on what metrics mean what; open rates, traffic, interactions, can all lead to conversion events.”
“One way you can approach this is campaign reports, instead of tactic reports. If views are lower than clicks, does there need to be testing on the page speed? Oftentimes clients will question the data in campaign reports as if it’s wrong - even if they receive a true conversion, how can they understand which metrics to improve on? With an aligned definition of data, the possibility of upsell is there; more work on the landing page optimization can go to your agency if your client understands the impact of the measurement.”
“For a data dictionary to be truly useful, data taxonomy has to permeate the agency and the client. How you define metrics, how you define the measurements and the calculations and formulas, all of that links back to value perception. If your client is looking to ascertain NPS from the dashboard, the calculations and events add up to actual bottom line understanding.”
How To Build Your Data Dictionary
With the importance of a data dictionary established, we can now work to identify the data sets, formats, and functions that are important to include in cataloging your agency’s data taxonomy process.
The purpose of a data dictionary, on both the agency and client side, is to avoid asking questions like, “what does this variable name mean?” and “what is the ideal value for this field?” This is why a data dictionary should include;
- Measurement units
- Synonyms and associated variables
- Names and descriptions of each variable
- A range of accepted values, including minimums and maximums
Just like brand guidelines, a data dictionary exists at the junction between an agency’s unique ability to create campaigns that align with a client’s brand, and an understanding of what performance metrics will be used by the client to gauge the success of those campaigns.
Every client has a unique business model, so the reporting requirements will be unique as well, but this does not mean the protocols for building marketing dashboards and reports can’t be templatized. It all starts with finding out how your clients measure success and what metrics are incentivized within their organization.
“A data dictionary is related to client KPIs and your agency’s “secret sauce,” says Dustin.
“How do you prove agency value through your services? This can only happen when you really understand what data, conversions, and metrics your clients value. The secret sauce of your agency is related to your goals and achieved/realized through the consistency of your reports. Brand style guides are no-brainers for an agency and a client; a data dictionary works in the same way for metrics and reports.”
A data dictionary acts as a reference guide on a dataset. According to Harvard Business Review, 80% of a data scientist’s time is spent finding, cleaning, and organizing data, leaving only 20% to perform analysis. That’s why a repository of all data assets — column descriptions, metrics, measurement units, and more — can help.
Since our main function is marketing performance reporting, the NinjaCat platform is engineered to take data from many sources and spreadsheets and connect/collate them into a data visualization format that then is used to build reports and dashboards. This is achieved through our Smart Connector setup process.
When bringing in data via a smart connector the data needs to be formatted in a column-based format. Depending on the smart connector that you are looking to use, these columns may exist in a raw data file (email, google sheets, FTP, S3, URL) or could also be in a database (SQL, BigQuery, etc.). For more information on data Smart Connectors within NinjaCat, check out our support article to understand what makes data a good fit for our platform.
Not only can a data dictionary ease an agency’s reporting headaches, it also can provide valuable strategic insights on how to build and measure campaigns.
“The agency can build a media plan on client-specified conversions,” explains Jen. “The kick-off calls should include conversations about what actions the client is looking for, and then the agency can create media plans that lead to those events. Reorientation is more expensive than initially having a strong onboarding process with reports.”
The reporting process is the lifeblood of the agency/client relationship. When you approach marketing reporting and data with a taxonomic and scalable mindset, you enable faster onboarding, empower rapid campaign creation, streamline your measurement protocols, and accelerate your agency’s ability to build your book of business.