Liquidity is "Easy," Valuation is Hard

The Provocateur: Darren Herman of Varick Media Management"The Provocateur" column is intended to incite discussion on a variety of topics around the evolution of digital media.

Darren Herman is Founder, Varick Media Management and Chief Digital Media Officer of kirshenbaum bond senecal + partners

Liquidity is "Easy," Valuation is Hard

- Current "demand side" technology platforms do not ingest client side data (other than ad-serving data) so most bidding is done for pricing efficiencies which is why publishers are freaking out.

- Unless pricing is done strategically by the true demand side, then the opportunity is limited for "media plan partners" as the sharing of the really useful data is limited.

- If we’re moving into a real-time environment, knowing what we want to pay (bid) for each impression is going to be key but a major technological undertaking at real true volume.

- Does pricing of inventory sit separately than inventory procurement? Is inventory procurement the "last mile" where "pricing" is extremely strategic?

Our industry is moving fast, but are we moving in the right direction and with the right infrastructure? This question is one that truly keeps me up at night.

There is a lot of focus in two areas of our ecosystem:

  • Media procurement
  • 3rd party data procurement

Are these the right areas to focus on first? Apparently the $100MM+ that has been invested in our space in the past 12 months, 100+ jobs created, thinks that this is correct. While I appreciate both of these areas of our ecosystem, I view both of these components as "the last mile."

Without naming names, many companies are building out pipes into inventory sources (exchanges seem to be the trend) to gain access to inventory pools to shift demand liquidity. Pipes in themselves are inherently dumb so the secret sauce that each of these companies are building is a platform that pings (for lack of a better word) multiple inventory sources to find the pricing inefficiencies and exploits them based on a rule set forth by the individual client (i.e. content rules, daypart rules, etc). No wonder publishers are freaking out.

What I’m spending quite a bit of time working on, and looking at, is the macro problem that needs to be solved for the long term: pricing individual impressions. While inventory sources such as Right Media, AdX, AdECN, AdNexus all tell us what the bid prices are for inventory, we, as marketers (agencies/brands), must know what we can and should pay in order to drive the right performance based on the campaign KPI.

I’m wrestling in my mind if the pricing of inventory should sit separately (but used in the ecosystem) of inventory procurement, as IMHO, pricing is extremely strategic whereas inventory procurement is looked at as "the last mile." Agencies will be out of business if they do not own the pricing mechanism in the long-run. Currently, they sit in the drivers seat here as they [should] have access to the majority of data that will make pricing predictions possible.

I see a world where the pricing mechanism ingests media and audience data (1st, 2nd, 3rd party – sorry IAB), business intelligence (NPV, etc), and provides real-time predictive bids, not just an exploitation of efficiently priced inventory. This can live totally separate of "the last mile." The major outcome of this is that key publishers for individual clients will start to see CPMs rise for their inventory because actual individual impressions can be valued and prices will rise for inventory.

If you think that this is already being done [at scale], I’d like to hear about it. If this was true, our media teams wouldn’t be making Nike sneakers after work or going out to Cipriani for lunch each day. I do not believe Madison Avenue could answer how much every single impression is valued which is a problem because hundreds of billions of dollars are spent there.

There might even be the "holy cow" moment when we figure out which part of our advertising spend is working for us.

Follow Darren Herman (@dherman76) and AdExchanger.com (@adexchanger) on Twitter.

12 Comments

  1. Great post. Two thoughts come to mind:

    He who holds the dollars needs to hold the (API) keys to the castle.

    If you have the right map there are always roads you can take that don't have tolls.

    Reply
  2. John Nardone

    Darren: With all due respect, the first line of your article is wrong. The [x+1] platform DOES ingest client data...and uses it to identify and value inventory. We do this for CPG clients through our partnership with IRI...and for clients like Delta Airlines, who own large customer databases.

    Reply
  3. Enjoyed the post. I think one of the key benefits of this first phase of building liquidity is lowering the very high transaction costs for online buying. We're just seeing the tip of the iceberg for smarter valuation.

    @John -- Has ingesting the client data been a game changer? What is the cutting edge of customer valuation? Are CLV calculations being entered into impression valuation?

    Reply
    • John Nardone

      Client data is a game changer is two ways...first, targeting based on client data allows us to leverage the clients insight, investments and research. We dont have to reinvent the wheel...but more importantly, we can and do model based on client definition of value...so we know what each impression is worth to each client.

      Reply
      • agreed 100%, enabling customer data in bidding is the 'end' goal of rtb in my opinion, reading value...

  4. Darren, you ask some very interesting questions, ones I have been trying to grapple with as well. As you pointed out, to me it seems the demand side is focussed on exploiting inefficiencies more so than using the true value of an impression (based on data - beyond the advertisers existing user base). Maybe its because its the lowest hanging fruit. I am hopeful that we would soon see startups that try to tackle this difficult problem.

    Reply
  5. Darren, what you've just highlighted here is that there is a big difference between a valuation engine (powered by advertiser data), and a standard DSP 'bidder'. The latter can make generic valuations based on freely available data points (e.g ad size, domain, category etc).
    The former relies on access to another much richer and deeper layer of 'advertiser' data (e.g previous purchase history, basket size, lifetime value), that allows a more bespoke, goal focused valuation to be calculated. This is the data that really matters, and the data that advertisers are most protective over (and rightly so). It's relatively easy to build or license a basic 'bidder' and access freely available or buy in 3rd party data. The smart buyers of the future (i shamelessly include Infectious Media in this group), are the ones who understand how to harvest,enrich and package advertiser data to power custom valuation and optimisation decisions at scale in real time. This is what will differentiate the buy side specialists, as it enables us to identify 'value' where others can't. On a more practical level, this means that any DSP or buying platform needs to be able to run different valuation algorithms for every advertiser and campaign KPI, based on the data points available. Some can do this already, and some can't 🙂

    Reply
  6. Darren,

    It is always great to read a provocative and thoughtful piece.

    I agree with you, it is a technology challenge for ad agencies to price each individual impression. Technology companies are working with the best agencies and advertisers, to automatically perform real-time media buys. They use newly developed high performance platforms that are able to simultaneously select, value and bid on each individual impression, tens of thousands of times per second.

    This fundamental change in the online advertiser market re-aligns the interests of publishers and advertisers. By doing an efficient pairing of content, consumer and creative, campaign performances are improved, and inventories are fairly monetized.

    Sandro

    Reply
  7. Darren - great piece, awesome insights as always.

    The holy grail is to be able to integrate into the advertiser's systems and to deliver the precise thing they are looking for - typically customers. Nobody should discount how difficult this is and will be to do in a scalable fashion because it means getting ingrained into an entirely more complex set of legacy infrastructures.

    Marketing-related software itself has typically been a fuzzier, softer, more disconnected set of systems anyway in the enterprise software realm, but this will change once customers see the benefit and have viable options to integrate the entire customer value chain.

    Moreover, the enterprise software giants if not already will soon be taking notice; I think agencies should tread carefully around "controlling customer data" to avoid waking the sleeping giants before they are really ready with their own, scalable technologies.

    Reply
  8. Nice post Darren, glad to see you agree with what we have been saying for the last year.

    As a side point, at Triggit we have always allowed customers to import their own data for targeting purposes, that is the whole point. We work with people's crm, website, conversion and user data to target and optimize.

    Reply

Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>