_ WHAT WE THINK

Insights

01

Questions to ask your agency about data.

Being a good partner is about asking good questions to help you and your partners.

Media reviews are not the easiest exercise. Depending on your scope and scale, it can mean months of back/forth, chem meetings, intros, deep dives, 1:1's and more. One topic sure to be a big part of the discussion is data. Agencies continue to evolve their capabilities around data and analytics. Some better than others. One key consideration is to ensure your strategy and interest around data. That is, today and looking into the future. In the best case, your culture is beginning to look at data to empower decisions. If not, that's an article I'm organizing as we speak.

Whether you're in review or not, data should be an ongoing topic of discussion. With that in mind, here are some topics to consider.

 

Strategy.

Do I want my media agency to build, organize and own my data ecosystem? Are there elements you are currently leveraging in-house? Are there external partners you're considering to help build your ecosystem.

 

Cost.

What elements of the data costs of my agency contract are fixed vs. variable, even if they’re pass thru costs? Rarely will the agency specific call out the technical debt in their system. Still, its important to know exactly what your costs are.

 

Security.

How is the agency’s data stack permissioned? Which laws, regulations and industry guidelines do they align to in the markets they manage, e.g. SOC2? Where does the agency’s data stack end, e.g. do they store Personal Identifiable Information (PII)? This is a very important topic and its important to know if they've had any breaches. If they have, what protocols are in place?

 

Sources.

What kinds of data sources, at what scale and at what granularity is data onboarded from the agency’s partners? It's important to get access at the most granular detail as possible. This can increase the size of data stored, but helps empower analysis in greater detail.

 

New Sources.

How do I get new data sources added to my agency-led ecosystem? Is there a cost to add new sources? This can be part of the cost conversation. Depending on the third-parties the agency uses in their stack, some charge by seat or by row of data or account, etc.

 

QA.

How is the agency reconciling data in the ecosystem? How are you able to QA data, especially planned vs. actual? Establishing a Single Source of Truth (SSOT) is key. The data integrity and ability to QA that data is even more critical.

 

Exports.

How do I get data out of my agency-led ecosystem? Whether its for internal modeling, external modeling, audits, etc. how is data exported? Can they do straight S3 bucket swaps or direct feed into your data lake? This is both for big asks and for small requests to follow-up on a signal.

 

Roadmap.

What is the future roadmap of features and functionality in their system? How often is the roadmap updated?

 

Access.

How do the agency’s team access data today, e.g. dedicated group/shared service, direct pulls by media team members, Jira-style ticket request system? The goal is to understand any potential roadblocks in advance.

 

Historical Data.

How do I transition my historical data and reports to the new agency?

 

Change + Transition.

What happens to my data and the ecosystem if I change agencies? As a brand, what are YOU going to do with a dump of data on your doorstep. What will the new partner do with it?

 

Outputs

How long does it take to get new outputs built, e.g. email report, export report, dashboard, etc.

02

Considerations about your data strategy.

Connecting the "how's" to the "what's" and "why's" among other considerations is key. Ensuring data is positioned to empower business is an ongoing dialogue. The dialogue itself can take on different forms. Here are some considerations to note as these conversations come to life.

Build vs. Buy.

This question has been debated since the dawn of technology. In recent years, the teams in marketing have found themselves in the middle of these conversations. Previously, this was an area that IT or Engineering teams did on their own but now with larger business impacts from these decisions, more seats are at the table. What's important is to not just consider the concept of build on it its own. Anyone can build something. Maintaining what is built is where things can go sideways. Especially with media data, API's change, fields come and go, connections across platforms evolve and technology moves. Building AND maintaining this by an internal team can be challenging. Also of note, the color of money in CAPEX discussions vs. how third-party partners are engaged. The total cost of ownership can change. Be sure all teams are away of the total cost, internal processes/procedures and all the "how's" in a system's operation. Plus, as your strategy changes, so might your system so building for flexibility is critical.

 

Objectivity +/- Subjectivity.

Leaders love the concept of automated insights. AI is getting us much closer to a world where optimizations are automatically introduced. That said, there is risk in the concept of having automated insights where other patterns in data might be missed. There is a balancing act among data objectivity, business/human context, market factors, perspective, etc. AI is an accelerant, however needs great inputs to see its full potential. The bottom line, don't forget about the narrative behind the data.

 

Data Integrity.

Garbage in. Garbage out. Hard to find a better example than with data. Ensuring the integrity of data is paramount to any system build. Questions around how data is QA'd, sample sets reviewed, hygiene completed and more all help to ensure that the data in your system is correct and dependable. A system without a focus on data integrity is doomed to fail.

 

Cadence.

We hear from partners about their needs for real-time data. What does that mean? What is "real-time". Whatever the answer is, a good reference point is to consider how you affect change. If you can get a signal, investigate, make the changes to creative+media+audience and see results in minutes, then you are real-time. Measure and consider optimizations at the cadence that best reflects how you operate. Aspirational goals are totally fine. Just know that how you affect change should connect to how you measure and report on data.

 

The Source of Truth.

The want for a single source of truth is a real goal for business, especially at scale. That said, it may not always be as clean as you want and that's perfectly ok. The single source may indeed be a series of sources that are connected and represented in a single output. The idea of hard-lining a single source of truth is not what it was in traditional IT/Engineering orgs, SDLC and related processes. The amount of data can be overwhelming. What is helpful is defining taxonomy, nomenclature and other ways of structuring data to ensure that regardless of the source, it will fit well into your system.

Welcome to Haystack.

We are your experts in building intelligence systems that bring data strategy to life for brands.

Let's talk about what you're working on and how we can provide some trusted, outside perspective.

Contact Us