Cost-Conscious Agencies Should Rethink the Real Price of Data

We did the math: managing a simple weekly report can add up to over a half-year of work.

By Cordie DePascale, Chief Strategy Officer, PremiumMedia360

Are you managing your company’s time resources effectively?

As we’re reaching the end of a pandemic that’s hit the ad world hard, many agencies are on the lookout for sources of “found” funds – pockets of operations and overhead where incremental efficiency gains can free up much-needed cash. To the CFOs and COO’s tackling this unenviable task, I suggest one place in particular to dive into: the cost of data. Or to be more precise: agencies should take a second look at the cost of managing data.

Since its inception, the media business has been a data business, with agencies and sellers executing and optimizing based on shared information. But even as advertising has become a new data hub in the “Mad Men to Math Men” revolution, agencies still struggle to get the data they need to make the decisions that matter. And that struggle is costly.

How costly? Some back-of-the-napkin math can be helpful here. And for the sake of example, I’d like to focus on just one specific subset of agency operations: preparing a definitive report comparing TV ad spots purchased with the ads that actually aired.

Creating that report – comparing agencies’ records of what spots have been guaranteed with sellers’ airing logs that register what ads have run – is a crucial part of any media buyer’s job. It’s this comparison that shows both sides what spots should be invoiced and what makegoods are required, telling both sides “what’s next” once the ad flight is complete.

But as crucial as this process is, it’s anything but simple – or quick. In fact, based on the basic “back-of-the-napkin” math below, a midsized agency can take over 25 hours per week, per client getting that data. I’ve outlined my thinking in the steps that follow – walking through the work a hypothetical agency faces for just one campaign of a single advertiser, running in 180 stations (representing a few dozen local TV markets) nationwide.

Getting Airing Logs from the Station or Network: 30 Minutes. For agencies, step one in auditing airing logs is just getting a hold of the logs themselves. To be sure, sellers are contractually obligated to deliver those logs (and their invoices as well). But even with those obligations in place, busy seller employees still need the occasional reminder to compile the reports and hit “send.” And all those reminders can add up: just one three-minute reminder call to ten sellers adds up to a half hour weekly – or two hours a month.

Converting PDFs to Usable Data: 2 Hours. Some (although not most) sellers send airing logs in PDF format, which most agency software can’t read. To import that data, media buyers must manually input the information. If just 8 sellers send airing logs via PDF and it takes 15 minutes to input the data on the PDF, that’s two hours wasted.

Automating Digital Data to Usable Data: 12 Hours. Most sellers send digital airing files – not PDFs – but those CSVs and files still take time to transpose into formats that agency systems can ingest. Sellers’ airing logs aren’t standardized, and sellers and agencies typically encode data differently, so media teams must “rewrite” seller airing files before they can upload them. The first leg in this transposing is running the files through automation that might take around four minutes per log file. For a campaign running across 180 sellers, that translates to 720 minutes, or 12 hours.

Data Cleanup: 12 Hours. Even with all that automation, there will still be kinks for media buyers to work out. In particular, since the specific commercials and the programs they run against change frequently, IT teams don’t typically have the resources to update the relevant ISCII codes, ad IDs, or program names for each round of automation. Instead, media buyers are left doing this manually, handling the “last mile” data wrangling of transposing that encoding into the agency systems. Assuming another 4 minutes of correcting per log across those same 180 sellers, that means yet another 12 hours.

By this point – with a total of 26.5 hours past – the airing log data are report-ready. They’re prepared enough that the agency can input the data in-system and find important information – such as spotting discrepancies with the seller’s version of the truth that needs to get resolved.

Again, this math is just an estimate, and the numbers won’t be exact. For instance: twelve hours spent waiting on automation means something different than twelve hours actively lost to manually adjusting codes. Meanwhile, none of this accounts for the surrounding time lost to misaligned data – including the hours the IT team spends writing automations, and hours spent negotiating makegoods because buyers and sellers aren’t working off a single version of the truth.  Meanwhile, just setting up airing logs data creates a whole new set of problems to be solved – including making sure you’re looking for the right data points, and you have a system and technology in place to find that data…although that’s a topic for another time.

That said, I’ve fact checked by running these estimates by more than one agency media group head, and they agree that my estimates are in line with today’s agency realities. They’re also in line with my personal experience from earlier in my career, when I ran data management in DRTV.

That 26.5 hours per week amounts to an astounding amount of time annually. Conservatively estimating a 48-week work year (assuming two weeks off for winter Holidays and other office closings), 26.5 hours a week is roughly 1,272 hours – or about 160 eight-hour days – per year. With around 20 working days a month, we’re talking eight months of a year dedicated to one type of weekly report for a single client.

Of course, managing airing logs is just one data operation agencies must conduct – among many. To be sure, there are steps agencies can take towards mitigating the data management problem, from basic checklists to more complex software. But whatever the approach, it’s crucial that agencies look to see faster data management as a critical goal. The time they save on data translates into found resources – at a time when they need those resources most.