If 2013 was the year of the selfie then 2013 was also the year of the self-serve in Media. We live in a DIY, customizable, controlled world and media buying’s self-serve trend is no contradiction to this new mentality. We create Nike shoes, Save the Date cards, and buy Facebook Ads all thanks to pretty user interfaces and unified payment systems (Paypal, Stripe, Braintree, Square). But is self-serve really that glamorous and cost effective for the power user/agency/trading desk/tech savvy client side marketing team? Joanna O’Connell beat me to the punch today and also is a bit more articulate than I, but here’s a blonde take on it:
Cutting the Agency Out…or not
One of the main arguments for self-serve is the presumption that by cutting out layers you cut your costs and drive up ROAS. What might be true if you look at a LUMAscape, but what we all know is not the ‘real world’ in media is that media doesn’t flow linearly and neither do markups. Let’s assume that Joe’s Shoes uses Media Guru Agency for their buying and does their creative work in-house. Let’s also assume that Media Guru Agency is part of a Biggy Big Holding Company. 50% of Joe’s Shoes goes through the Biggy Big’s trading desk, 30% goes to ad networks, and the last 20% goes to direct buys. I’ll let you do the math on the arbitrage here, but let’s just look at straight operational costs for leveraging self-serve in-house:
Assuming the agency takes a 15% markup on media spend Joe’s Shoes $250,000 Q1FY14 budget is planned for $212,500. Joe’s Shoes pays Media Guru Agency $37,500 quarterly or $150,000 annual to plan and execute their campaigns. For Joe’s Shoes to take this in-house it would be more expensive by looking at salaried employees (this does not account for benefits, free shuttles, and foosball):
Director, Programmatic Buying ~$110,000
Programmatic Display Manager ~$95,000
Paid Social Manager ~$85,000
Partnerships/Direct Buys ~$95,000
Marketing Business Analyst/Excel Monkey ~$90,000
Ad Ops Specialist ~$80,000
Joe’s Shoes would employ a team costing them $555,000 annually to do what their agency did for 70% less.
Workflow Shouldn’t Be Unified
Speaking of agencies, what some don’t realize is that self-serve media buying is primarily used by two totally different parties: Agency Trading Desks and SMBs. One services some of the largest agencies in the world with buying power even Bitcoin holders would find impressive while the other works scrappy and sends emails to an alias for buying support. Yet self-serve products try to accommodate both parties needs. One of the best examples of self-serve’s evolution to become less mass accommodating and more focused is the Facebook Initiative workflow roll out. Facebook took a step back to improve workflow and instead of everyone seeing the same bells and whistles they created a more streamlined and customized workflow depending on if the advertiser wanted reach, acquisitions, app installs, etc.
Features that aren’t important to the power user (Agencies) such as stock images, off the shelf reporting dashboards and standard attribution models are things SMBs need. Agencies looking for data export features, robust targeting and optimization functionality, and congruency with existing platforms they’re using (such as DMPs) can create product pipeline headaches. Moreover, while one would argue that self-serve should accommodate the power user first, SMBs collectively make up a massive amount of media spend (~$1.5BN in Paid Social alone). So where do you put your efforts? You could do what Facebook or even Google did and offer self-serve, but create an ecosystem of platform and vertical specialists (the Salesforces, Marins, Nanigans, of the world) that can be an advertiser facing layer that helps fully realize an advertiser’s spend. So now we’re creating another middleman by offering self-serve? Not really, just changing the way the advertiser interacts with the middleman. Instead of executioners they’re now specialists and platform advisors.
The Middleman Becomes the Specialist
I bet the majority of the volume going through self-serve today is for retargeting. Basic retargeting is like the Easy Bake Oven of self-serve tactics. Doesn’t require adult supervision. Advance to the Viking Gas Range with Prospecting and more dynamic retargeting tactics to realize the value of the ‘middlemen’. Self-serve across other mediums - connected TV, video, mobile, etc. are somewhat negatively impacted by self-serve because testing within self-serve is less frequent without the educational support for emerging media. Clients push agencies to think differently, test and trading desks should push their self-serve in the same manner. However, self-serve is just a tool - not a recommendation engine. Until it becomes that way, agencies need to leverage the self-serve provider as specialists into all inventories, platforms, and formats they access. Take your DSP for instance, the DSP sees the whole universe you bid through and therefore can give you a macro view of the marketplace dynamics, trends, targeting parameters and nuances of each channel (mobile, video, display, etc.) that it includes. What does this mean? DSPs should staff up on planners and people able to advise power users from a place of authority rather than be passive and just there to troubleshoot. Agencies can work with self-serve providers to make them bring more strategic offerings to them and their clients - private deals, data partnerships, alpha programs with Facebook/Google/inventory sources and best practices.
If self-serve is your ticket to cutting costs then think about the executional layer, billing, and technical aspects before you ditch your agency. If you’re providing the self-serve services - create a consulting layer. Not just a troubleshooting layer. Ok, now go read Joanna’s post because it’ll add additional substance to my position.