A Call For Sharper Bidding

By
  • Facebook
  • Google Plus
  • Twitter
  • LinkedIn

casale-ddt"Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today’s column is written by Andrew Casale, vice president of strategy at Casale Media Inc.

It goes without saying that increased familiarity with programmatic buying has brought forth a lot of great executions among marketers.

Today we see real creativity and sophistication in the way media is bought, where marketers know what they want and technology has caught up to allow them to find it quickly at scale. This is no more evident than at this time of year, when we see literally thousands of new executions launch daily, capitalizing on the all-important holiday period.

However, with all of these sophisticated tactics now available -- i.e. marketers can buy the auto intender who just left their site, who spent weeks researching a vehicle, on the device, in the environment and during the time frame most conducive to influence -- there’s one incredibly important area where marketers and media buyers still fall down and generate waste: the bidding process.

In RTB, buyers know the value of their target audiences, but only a few have prioritized placing an optimal value on the media itself. Technology has enhanced the process of buying, but bid prices more often than not still seem subjective and human-defined.

When we analyze seat-level bid pricing strategies, we see wide variance in methodologies and subsequent outcomes. And all too often it feels the strategy may just be, “You know, a dollar-fifty sounds about right.” When this happens, the value of an incredibly sophisticated channel goes out the window. Today the worth of an impression is defined by demand in the market, more so than floor prices.

Media buyers should understand who they’re competing against when they place a bid in a DSP. They know their intended audience, and they set their parameters to find it. Often, they expect that when they set their bid prices, they’ll be trying to beat publishers’ floor pricing strategies — which were the prices to beat in the earlier days.

But today, algorithms in many platforms have been tuned to locate the finest impressions in the market, and competing brands end up gunning for those same audiences. In this paradigm, buyers aren’t competing against publishers so much as they’re competing against other brands. This leads to the scenario today where on average more than 90% of bids placed are lost, and in the majority of those cases the reason for loss is competing bids, not publisher floors.

These odds aren’t great. Obviously it’s not in the buyer’s best interest to lose almost all the time. But it’s also not a cause for despair. It’s a wake-up call to bid more strategically. After all, there are also buyers that we see across the exchange who win 90% of the time.

The remarkable thing that we also observe in these cases is that when we compare the strategy and outcome of a buyer who loses 90% of the time to a buyer who wins 90% of the time, the clear price isn’t all that different. This is a function of the second price auction that the sell side operates. A winner only pays the second-highest bid price, and if that is priced at a significant discount which often it is, the winner tends to get a really good deal.

Buyers should think more about optimizing to the second-highest bid, instead of their opening bid price, as this is effectively what becomes their clear price. If you force yourself to bid no more than $2 because that is the ceiling you are willing to pay, you’ll limit yourself from a multitude of scenarios where a $3, $5 or $10 bid may still clear far below your $2 ceiling. Buyers should trust the technology to do its job in the bidding process, and that it will deliver the value it promises. The marketplace will sell all of the best impressions out there. If a bidder doesn’t understand the value of the impressions they’re buying, they keep simply bidding into the hands of those players who define the value of the marketplace.

Let’s mint more programmatic winners! You’ll find far more success if you’re winning 90% of the impressions that you have painstakingly identified as your optimal audiences. And if you’re winning them, you can take comfort in the fact that your competitors aren’t.

Follow Casale Media (@casalemedia) and AdExchanger (@adexchanger) on Twitter.

  • Facebook
  • Google Plus
  • Twitter
  • LinkedIn

Email This Post Email This Post

By on at

4 Responses to “A Call For Sharper Bidding”


  1. Shuai Yuan says:

    >If you force yourself to bid no more than $2 because that is the ceiling you are willing to pay, you’ll limit yourself from a multitude of scenarios where a $3, $5 or $10 bid may still clear far below your $2 ceiling.

    Didn't follow this. So paying CPM $10 is OK when you have a limit at $2? This is only true when you have no budget constraint. Bidding at the price that you can afford is the key advantage (truth telling) of 2nd price auction (even though it is usually screwed up in online advertising). You can lower the bid to fight against the soft floors, but increasing it would do little good.

    >You’ll find far more success if you’re winning 90% of the impressions that you have painstakingly identified as your optimal audiences.

    Again, this implies you have basically not budget constraint. The optimisation goal is "performance" (cvr or convs within the budget) but not "volume". An analogy: always winning the Rank 1.0 position and Rank 1.1 in AdWords result in great difference of cost but much less in performance. It is not about winning the majority of impressions even they are perfectly "fit" --- you always need to have a bid landscape curve in mind. Optimisation without budget consideration is not helpful.

  2. George Bouras says:

    I'm not sure I follow...

    "If you force yourself to bid no more than $2 because that is the ceiling you are willing to pay, you’ll limit yourself from a multitude of scenarios where a $3, $5 or $10 bid may still clear far below your $2 ceiling."

    The impressions which clear for lower than $2 would also be won with a $2 bid; you don't need a $3, $5 or $10 bid to win them at those prices.

    I thought the beauty of the second price auction in RTB was that it encouraged a buyer's strategy to be purely about the value of the impression to them, freeing them from having to worry about the bids of rival bidders (putting aside worries about securing a minimum volume).

    Is there something I'm missing?

  3. Andrew Casale says:

    You're both right.

    The point I am trying to raise is that it does not appear that the common bidding model implemented is a buyer's true valuation of an impression's worth. Instead, the prices feel arbitrary, likely driven by attempts to beat fixed floors, or a fear that a high bid will always clear high. This may have helped get a better clear price when density was low, but today it largely leads to heavy losses.

    The ultimate point I am trying to raise is for buyers not to fear bidding "too high" as rarely do they clear anywhere near that high, and by not taking that risk they are ceding a huge advantage away to those that do.

  4. Paul Denver says:

    I agree with the previous 2 comments - there should not be any impressions that 1) you can win with a bid of say $10 and 2) pay less than $2, that you wouldn't also win by bidding $2, and you would pay the same.

    That suggests that the differentiating factor between winning 10% of impressions or 90% of impressions is not necessarily bid price, but another factor. I'd suggest it is likely to be competition. You say there is lots of competition overall, but this won't be uniformly distributed. For example, if there is a lot of competition then winning 10% of available impressions isn't too surprising. But if you are able to win 90% then there likely isn't much serious competition for that particular set of impressions.

Leave a Reply