Why An Independent Buy-Side And A Sell-Side Makes Sense…For Both Sides

By: Joe Zawadzki  Published: January 11, 2012

Link to original article

There has been a lot of discussion of late about convergence between the buy and sell side – with a number of large players trying to build their own end-to-end stack. Here Joe Zawadzki, CEO at Mediamath, argues that an independent buy and sell side is in the best interests of the industry.

There is an interesting question being posed in ad tech right now: should the buy side and sell side be allowed to coexist within the same platform? Put another way, how much does focus and loyalty to a specific class of customer matter?

It turns out, a lot, for at least three big reasons:


In finance, playing both sides is what precipitated the financial crisis of 2009. Investment banks that had CDO factoring capabilities as well as retail clients to buy them – buy-side and sell-side – were at the heart.

They knew what clients wanted and how much they had to spend, their risk tolerances and rating systems.

Investment bank Team B used the insight and client access from Team A to produce something that Team A could sell as an agent that was wildly profitable for Team B. Human nature took over as Team B took more and more margin – incentives matter and Hamptons houses don’t buy themselves – by putting riskier and riskier products into the hands of Team A to sell until … read “The Magnetar Trade” by Jesse Eisenger for a good summary of the outcome.

Playing both sides creates conflict of interest, where the profits of the arbitrageur or volume-based aggregator are created by withholding value from the principals on both sides.


It’s a truism that “strategy is what you don’t do.” In any organization, no matter how large, you can only be truly great at a handful of things. Trying to be great for marketers, for trading desks, for intermediaries, for portals, and for publishers is impossible. Even the largest firms tend to struggle – Google has struggled with social; Facebook’s phone is delayed.

As they say, jacks of all trades are masters of none.

Beauty is in the Eye of the Beholder.

Here’s the subtlest-yet-most-fundamental ah-ha about buying an impression relative to buying other things: an impression has no intrinsic value for all buyers, and is instantaneously perishable.

What does that mean?

An impression in front of a novelist dad who is actively looking for gifts for young kids from home who is a regular e-commerce shopper on December 12th is worth a lot to Nintendo and Mattel, and perhaps nothing to Cisco. Whereas the next impression to the single CIO of a mid-sized tech firm might find the value of those respective impressions inverted.

Thus, the ability to determine, and communicate, what an impression is worth to a specific buyer, or specific offer, or specific creative is the rosetta stone to marketing.

If the platform you, as a buyer, is using can’t value impressions appropriately because it doesn’t have the focus, doesn’t have the technology, or is rife with conflict, you can’t communicate value appropriately.

If you, as a publisher, can’t glean true value from your buyers you need to offer up average prices for average impressions, which is a bad business. The 50-millisecond moment passes and a random ad is jammed in the space at truly remnant prices.

They say you “can’t manage what you can’t measure.” I’d contend, “marketers can’t pay up for what they can’t predict,” which ultimately hurts the publisher.

To put it in a heavier-handed way: DSPs help publishers by letting marketers accurately bid based on true value – everyone finds their best 5% and pays for it, as opposed to having to buy everyone and factoring in the lossiness of the 95% out-of-market in their average price.

Let’s talk real numbers. Let’s say a marketer chooses a DSP that has 10% less supply available because of conflict and is just 10% less effective at optimizing marketer results due to split focus. With a $50MM digital budget across display, video, mobile, and social you’ve made a decision that cost $10MM.

If you’re an agency using a sub-optimal DSP, and you’re trying to win business, deliver great results for your client, and retain business while competing with an agency that uses superior technology, it’s an equivalent mistake. Search agencies that license Marin or Kenshoo fare better than those that use worse platforms (or manage search “by hand”). Financial firms that use Bloomberg and the latest quantitative models do better than those who use e-Trade and Excel.


Let publisher agents/technologies optimize for publishers. Let advertiser agents/DSPs optimize for buyers. By working together with transparent incentives, both sides actually win.