By Jim Head, President, Fitzco Analytics
For the past 20 years, I’ve lived many lives in the agency analytics world. I began at an independent agency. I was embedded within a big network agency. I started a practice of my own, working directly with clients and white labeling with agencies. And now, I’ve boomeranged back to agency life – having sold my practice and running it within the walls of Fitzco since late 2021.
An important part of my job is bias , that is, reducing people’s perception/agenda to form conclusions in favor of using data. For our clients we do exactly that. But as humans, we can’t eliminate it entirely, so even I have certain biases about the rise of agency “analytics” – the quotes may be a bit of a giveaway – that no amount of scrubbing can remove.
At the center of this is an often misunderstood distinction: analytics as general analysis and reporting (e.g., media performance dashboards) vs. analytics as contemporary data science (e.g., classification modeling, optimization, etc.).
My position is not to rank one ahead of the other, only to point out that there’s a difference that requires unique skillsets. How and with whom an agency has built its practice makes all the difference.
If it’s your job to discern whether your agency partner is built for the former or the latter, I don’t envy you. Mismatching the scope and depth of your needs with their practice capabilities can result in outcomes ranging from immaterial to wasteful to malpractice.
A few questions, observations and insights will help you know exactly whom you are working with. Distinctions fall into three categories: capability, technology and culture.
Capability. Knowing if your agency partner has these analytics capabilities will help you determine if they are the appropriate resource for your critical marketing measurement, budget optimization and media targeting needs. (Caution signs are in italics.)
- Does your partner employ a full-time applied statistician with graduate-level coursework and programming capabilities?
“We have a contractor that can help us with the heavy-lifting when it comes to statistics and modeling.” - Can your partner distinguish the theoretical outcomes of their work versus the contribution they make to your stated objective(s)?
“Our models are 98% accurate and we work with a professor at a local university to validate our findings.” - Is your partner able to describe, in plain language, the methods they use to answer different questions (e.g., a time-series analysis, a classification model, a segmentation algorithm)?
“We have a proprietary tool just for that.” - Will your partner contemplate and include data sets aside from those you provide to them? Are they conversant with publicly or commercially available data assets to complement those provided to them?
“We will load the data you send to us into our A.I. platform and go from there.”
Technology. Capability can overcome technology in some cases. But an analytics stack speaks to how efficient and comprehensive your partner can be. Knowing if your partner licenses these tools, programs and environments will help you determine where they lie on the analytics continuum. Again, italics signal caution.
- Does your partner require code-based solutions for many/most of their workstreams?
“We favor programs that automate code because it allows us to deliver faster.” - Are platforms, packages and languages like Python™, R/RStudio®, SAS®, SQL a regular part of your partner’s vernacular?
“We find that spreadsheets, Tableau® and Google Analytics™ meet most of our client’s needs.” - Are tools selected because of features unique to the needs of the engagement?
“Our senior analysts only work in Power BI™.”
Culture. The collective personality of your agency partner’s staff is, perhaps, the best indication of whether they are a seasoned analytics practice or a pool of marketing analysts. An experienced analytics team will listen when consultants present as story tellers, coordinate as project managers and incorporate feedback without ego or presumption. Observing these characteristics will tell you a lot about whom you are working with. (Note italics.)
- Does your partner encourage a thorough pre-engagement briefing with a problem statement, an objective and a quantitative description of success?
“We know what is wrong with your business; we will be back in touch soon with the solution.” - Is your partner practiced in translating complex statistical outcomes into business-friendly, practical and deployable solutions?
“This presentation summarizes our findings – we’ll let you take it from here.” - Can your partner absorb and incorporate business-critical feedback that does not align with their favored approach?
“We are the experts in this area – deviating from our solution is not advisable.”
All these considerations—capability, technology and culture, really need to be assessed by marketers before any partnership proceeds.
