I’ve been studying ad agencies for the better part of 10 years.  First as an insider, working in media planning and analytics at a full service agency, and now as an outsider working for an advertising technology company.  Over the years I’ve learned a lot about agencies, but to this day I continue to learn new things about them and the critical function they play in the advertising value chain.

One of the most interesting things I’ve learned lately is how difficult it is to sell technology to ad agencies.  At first glance, agencies seem to be the perfect candidate to adopt new technologies.  Pretty much everyone agrees that the current media agency workflow is inefficient and agencies feel lots of pain because of the manual nature of much of their work.  Why then, is it so difficult to sell tech to agencies?

First – let’s explore why agencies need technology in the first place.

The truth is: ad agencies haven’t always needed advanced technology.  When I was joining the agency world in the mid 2000’s – we barely had working desktop computers.  My first computer as a media planner was from 2001, ran an outdated version of Windows and was completely locked down by our central IT department.  Since we didn’t have admin rights to our computers, we couldn’t even install an updated version of an Internet browser without submitting a formal request and waiting several days for IT to come around and install our desired software.  Also, to save money on software licenses, we had one shared computer that ran our most advanced technology: a program called AdViews, which we used to access Nielsen’s database of global ad spend – allowing us to evaluate the media spend of our advertisers vs. their competitors.  AdViews reports took 3-6 hours on average to run, and, personally, it took me 3-4 attempts before I got the data back that I wanted.  Everyone knew that our work wasn’t very efficient, but the culture was very oriented toward solving problems with duct tape and man-hours rather than employing new technology.  Our work was also very cyclical.  Back then “planning season” encompassed the five months between May and September, and there were weeks at a time when people had no work to do at all.  The system actually worked pretty well until digital media began to disrupt the way agencies worked and operated.

The thing about digital media that makes it fundamentally different from TV, radio, print, and OOH is that the transaction volume is much higher.  For traditional media channels, the majority of buying is done on an upfront basis, where one transaction (e.g. during TV “upfronts” in June) can purchase an entire year’s worth of media.  However with digital media the transaction cadence is much different.  For agencies that buy “direct” using traditional means there can be as many as 4-12 transactions per year (when agencies issue digital IO’s) depending on the planning and optimization cycle.  However, for agencies buying through programmatic means, the transaction volume can be in the millions per day.

Here, the amount of technology required in the agency is proportional to the number of transactions that the agency has to handle.  Doing one upfront transaction per year requires a couple of Excel spreadsheets, a telephone, and a fax machine.  However, 100 million transactions per day requires super advanced, high throughput, low latency bidding and optimization technology.  In the past 10 years agencies have had to come to terms with the fact that they need access to this new technology in order to keep up with the order volumes necessary to survive in the digital world.

So why is it so hard for agencies to adapt new technology?

Many people say it’s because agencies don’t make enough money to invest in technology.  I actually don’t agree with this opinion.  It’s true that agencies have experienced significant downward pricing pressure from advertisers, but even if agencies made lots of money I still think agencies would experience the same problem with adopting technology.  The main problem here, I believe, lies at the center of the agency value proposition: flexibility.

More than anything else, the main value that agencies provide to the advertising value chain is flexibility.  They can morph and manipulate themselves to accommodate any new advertiser requirement and they can change very quickly if they absolutely need to.  Most change is spurred when an RFI comes in to an agency from an advertiser with a new requirement that the agency must fulfill in order to be eligible for the advertiser’s business.  When this happens, agencies instantly change shape to accommodate the new requirement.

 

As an example, I remember when media analytics became the must-have agency asset and an “analytics capability” all the sudden appeared on every advertiser RFI.  Seemingly overnight, all agencies spun up brand new analytics capabilities.  The formula was actually relatively simple – and we repeated it in several different areas.  First we hired one VP level outsider who was an expert in analytics.  Then we took 2-3 of the smartest people from other parts of the agency and paired them with the new hire.  In the agency where I worked, we pitched our new analytics capability as a key differentiator before the new VP had even found their way to their desk.  Our flexibility is what allowed us to stay on the competitive edge in the agency world.

 

Now for the rub.  The hardest people to sell technology to are those who need to maintain a constant state of flexibility.

 

Technology sales, especially sales of new technology, require customers to commit to the product they want and the price they will pay.  For example, if a customer says they will pay $1 million for a new technology over 10 years, then the technology company knows how much they can safely invest in people and resources to develop the product.  In the absence of getting a customer commitment it’s extremely risky to develop new software because there’s a good chance no one will ever use it.  In the absence of getting at least one customer to sign up to buy or use a product in advance of building it, there is probably somewhere north of a 95% chance that your product will fail.

 

This is why it’s so hard to sell technology to ad agencies.  In order to maintain their value proposition, agencies must maintain their flexibility.  However technology companies need customer commitments in order to justify the extremely expensive product development process.  It’s definitely not a lost cause – but it’s a challenge that many have been unable to overcome.

Why It’s So Difficult to Sell Technology to Ad Agencies
Tagged on:         
  • Yeah, basically.

    Agencies are just – I think I’ve said this before – glorified, industry-specific outsourcing firms. A look at the income statement of any of the major ad agencies supports this – they are basically labor + overhead, with a(n average) 10% margin.

    No agency can ever commit to any kind of long-term planning, because they always respond to customers. So, they are always behind. Always responsive.

    One problem is that an agency wants to deliver what’s in the scope of work. And – that’s it. There’s no reward for going significantly beyond the basics, and there’s no success lottery for a handful of really successful companies.

    The bigger challenge is technology companies cutting agencies out of the process. It makes sense for a tech company to sell to an agency (as a middleman) if – and only if – they can negotiate a handful of deals (say, 10) and cover most of the market; agencies can then sell the technology company to existing clients (say, hundreds).

    But if agencies do a poor job on the sell-though for a good product, then the tech company may build out a sales force and sell directly to end customers. If the product is sufficiently good, the end customer will buy.

    The more things can be automated, the greater the advantage that a tech-centric company (build product/service, sell that product/service) has over an agency (staffing model). If tech companies automate customer onboarding (see: Amazon Web Services), then you may actually end up with a large number of disaggregated, specialized tech firms that serve a customer instead of large conglomerates.

    See: Coase on transaction costs when in-market transaction costs drop to a low enough level.

    Still, that’s pretty far away.

  • I think you pretty much nailed it here.