By Adam Faughnan
Director of Offline Marketing, Credit Karma
The ‘it’ was January 2016 marketing spend and usually for a marketer this would be positive news. Especially for one who has worked at startups where the burn rate could only be higher if we were literally burning cash to heat the building. The sense of dread was more related to the fact that it was December 16th and spending an additional $10 million in a month while measuring ROI was about as easy as working on the Pepsi social team after Kendall Jenner healed the racial divide in America with a commercial. But, I stayed stoic and responded calmly to the CMO that it should be doable as long as we were able to relax our measurement requirements.
I quickly went to work. The goals were aggressive. The marginal efficiency had to stay within a tight range. But we had never spent at this level in one month before. It is always fun to spend money (my wife will attest to this) but less fun when you are traveling to Ireland in eight days and you need to ramp up multiple tests while figuring out ways of measuring them. I asked our TV agency for some scenarios. They had been planning January for the last month. At various levels of spend, what would we expect to be the difference in reach and frequency? What would the separation be between spots on each network? What creative rotation should we use to maximize response? What is the optimal mix of 15 and 30 second TV spots? What additional networks or tests would we consider at different spend levels? To their credit, they moved very fast, worked a lot of late nights and came back with a multitude of options.
Then came the fun part. I will admit I am not the best at responding to every sales email I get (tip for salespeople- don’t send the exact same email to the CEO, CMO, VP of Marketing and me and expect a response). Now I could email vendors and ask for quotes, but only if they got back to me within 24 hours with a clear proposal, dollar investment and the lowest CPM they could offer. Things that seemed a little crazy a few months ago (hello movie theater advertising!) suddenly had traction because you could get a fire-sale CPM, Nielsen ratings and do hold-out markets locally to try and determine ROI. Obviously, every test wouldn’t turn out ROI positive but the “what do you have to believe?” had to be there. Is it possible, with this target audience and this CPM, that you would get a response rate that would be ROI positive, given all historical testing?
Tests that were being mulled for Q2 of the following year were quickly moved up. The target was a younger demographic. Cord-cutters (selfish millennials who I suspect were still using their parents’ cable log-in) were flocking to Hulu and YouTube. Committing large investment amounts during a seasonally weaker period for advertising allowed us to secure rates that were usually not possible. Again, hold-out markets were used to measure response. While typically I would focus on local match market tests to accurately measure incremental response and marginal cost, the goal was now to drive as much acquisition as possible while trying to get an understanding of response.
Podcasts and YouTube endorsements were quickly ramped up and measured using vanity URLs. While the attribution multiplier was still slightly murky (like a muddy puddle), measuring podcasts against each other for efficient response would enable future testing. The final piece of the puzzle was an attribution survey we had been running on our registration flow that was quickly updated to include additional testing options. With our high level of traffic, we could establish some baselines in December before measuring changes in January once we turned the media on. All media tests were lined up, markets allocated, invoices signed, legal approvals procured in eight frenzied days. Then, for a real challenge, I took three kids on a ten-hour flight (two of them on laps) before the age when you should shove an iPad in front of them and tell them you would see them in Dublin. Who says offline marketing isn’t exciting?
– Testing: Despite a lack of lead time and a broad goal of crushing the acquisition forecast, it is still important to set up tests in a (SMART) way – specific, measurable, achievable, relevant and time-bound.
– Pricing: It can sometimes work in your favor to approach media companies at the end of the quarter with a low rate proposal if they have unsold inventory they are looking to shift.
– Attribution: Testing multiple platforms simultaneously can lead to difficult attribution but can also give you insights that will enable future tests (Hulu showed enough promise to launch a future test later in Q1).
– Ubiquity: Appearing everywhere in a defined period of time can be expensive but does have its benefits. By ramping spend in TV, Radio, Podcasts, YouTube, Hulu and Movie Theaters, we experienced an up-weight effect for our overall response for both Offline and Digital