Monday, November 30, 2009

Six Supply Side Strategies for Ad Networks in the New Ad Ecosystem

In my last blog I covered some concepts around the supply side of online advertising. In this blog I'll be venturing over to the demand side.

Ad networks have always been a two way street. The yin and yang of the demand and supply sides of the better ad networks growing and feeding each others needs in a virtuous cycle. As most online publishers, advertisers and ad agencies know, some ad networks are better at one side or the other, but over time a balance is found or growth is stunted.

Ad exchanges are going to be a game changer for ad networks because they break the need for a complete dependence on the other side of your network. And soon, when they start massively scaling, they also will have a profound effect on strategies for targeting and optimization and how campaigns are sold and filled. This effect will ripple through to publishers, advertisers and ad agencies.

Additionally the emergence of demand side only platforms acting as buying agents for agencies offers a disintermediation risk.

I am writing this blog entry from an ad network management perspective so if you are an ad network, here are some strategies for you to survive the change, and if you are a supply or demand side only platform here are some opportunities for you to explore. I'll have some new strategies for this list in the future, but here are a few starter strategies to wet the whistle.

1. Step Up Your Game...NOW!
You are now or will soon be competing with hungry startups who care not about balancing the distribution of budget and engineering resources between the supply and demand side because they will be demand side only or supply side only platforms. They will innovate faster, they will pay more attention to the advertisers, agencies or publisher needs, they will invade niches, exploit your weaknesses, and they will clean your clock. Invest now. Right now. Seriously, right now. Take some of that cash you've been squirreling away and invest it in your technology and your people.

2. Parallel Experimentation is Your Friend
Does your system support multiple optimization strategies in parallel? Can you experiment on the fly? Can you do post hoc testing and make sense of the results? You need to have multiple parallel experimentation streams that are constantly informing your optimization secret sauce and evolving against each other in an algorithmic survival of the fittest. Remember, optimization will no longer be just amongst sites or ad units, but amongst entire purchasing platforms!

3. Take the Channel Conflict Lemon and Turn it Into Margaritas
Start thinking about your sales team differently. In the old model the same person was your starting pitcher, your middle reliever and your closer. Now you are going to be using them very strategically. Your sales team actually becomes MORE important because the campaigns for which you will be using them most with be the ones that are more strategic and require high touch, hands on management. Sure, you will be buying inventory algorithmically (a.k.a. programmatically) for more of your campaigns in the future, but the winners in the new paradigm will be the companies who know how to balance that opportunity with a highly skilled group of sales people for the management of the more complicated deals and relationships. More on this in a moment when we talk about the leaner meaner sales team.

3. Create Your Leaner, Meaner Sales Team
If you are an ad network then no doubt you will have been keeping an eye on cost of sales as one of your key metrics. You need to now start thinking differently about that metric. What is important now is OVERALL cost of sales AND cost of sales per campaign. You should start looking at managing campaigns as either programmatic intensive campaigns or sales intensive campaigns. What you should be doing is paying a leaner meaner sales team higher compensation and getting your reduction in overall cost of sales on the campaigns that work smoothly in a programmatic model without high sales team touch. Don't be afraid to let the variation of cost of sales per campaign increase greatly (it actually should), so long as the overall cost of sales comes down over time. Don't get greedy (yeah right...) and start leaning on cutting the cost of sales for those high cost campaigns. So long as the overall cost of sales is coming down, then the higher cost of sales for some campaigns is an indication that the strategy is working. Think long term relationships. Some campaigns need sales people and good sales people get good and stay good by being supported, developed and well compensated. If you keep the advertisers and agencies happy on those high touch campaigns the programmatic campaigns will follow.

4. ROI is Your New Religion
If you are an ad network you probably review and approve the CPM (or other pricing model) per ad unit type your sales team is selling. Maybe you have a rate card minimum to keep your publishers in the fold and improve overall revenue. You certainly view ROI for your advertisers as important. Now you must make ROI your religion. ROI for your advertisers or the agency campaigns is now your guiding principle and should drive your product solutions. You need to understand ROI deeply because you need to know how to bid, fill the campaign and still make a margin. And how do you come to a deeper understanding of ROI expectations?...

5. IT Introductions
How are you measuring ROI? How are your advertisers or agency partners measuring ROI? How do they want the data? Do they want to API into your system to feed their system? Do they want a fantastic UI and canned reports? Do they want a UI that can query your data set and provide flex and pivot reports? Are you working well as the go between third party ad server data and the final user of that data? Is your internal reporting system able to easily incorporate data from multiple exchange platforms? Are your supply side partners, either in-network or third-party, aware of your forecasted needs in their systems. Think about how successful Walmart has been by integrating their POS with their suppliers. Use that lesson to guide development of your systems. Now is the time to introduce your IT department to outside party IT departments and help them work together. This may be a paradigm shift for your IT department. You may need to change the character of the team to accommodate exposing them to a team different in culture and with different goals and expectations.

6. Know How to Value the Impression
If your ad network has been going down the behavioral targeting path for some time, you have been moving toward targeting and optimization at the impression level. Now prepare to bid (or offer), in real time, programmatically in such a way that you are maximizing ROI for your demand side partners while lowering overall cost of sales. You do this and beat your competitors by being better at knowing how to value the impression. You should be capturing every available shred of data about the user, the site, the ad unit, the time of day, the season, special events, etc.. and using it in your feedback loop. You should prepare to incorporate third party data into your mix. You should prepare to evaluate the cost of third party data at the impression level in the final analysis of your margin. The bottom line in a bidding environment, if you know more, you can bid higher, win the auction and still retain margin. Strategy and real time execution of your strategy programmatically will empower your drive to maximizing ROI for your demand side partners.

Fair Use Policy: Please include the name: Edward Boettcher and the following link: The Product Messenger when referencing this information. All content (c) 2009 Edward Boettcher

Thursday, October 22, 2009

Forecasting and Yield Optimization

Large scale publishers, ad networks and similar such online inventory aggregators have stepped up their quest for yield optimization and revenue maximization. I've seen this before, so allow me step back in time...way back (in Internet terms) to the year 1998. I was a few years out of college. After a brief stint in the Finance Industry working on a trading desk at a broker dealer, I decided that my future lay in building businesses. I took a position in inventory and revenue forecasting at the Vermont Teddy Bear Company where I learned how to program and build databases. Shortly thereafter I was recruited by Sara Lee-Champion Jogbra where I really dug into the statistical aspect of forecasting. In 1998 statistical forecasting techniques were just beginning to get attention as a potential profit optimizer for the company. Using some data mining techniques and the statistical forecasting engine, Forecast Pro, I realized we were making the wrong ratio of bra sizes, particularly when you look at the size ratio by style, for instance a more supportive bra with an underwire ran to the larger sizes, where a compression style ran to the smaller sizes (yes I am an expert in both teddy bears and sports bras.) Given that sports bra design and colors for a segment of the product line were seasonal decreased the lead time for making adjustments to production. Additionally, unsold seasonal sports bras qualified for return to manufacturer where we would reroute them to bargain outlets thus hitting our margins (ahhh yes...the direct correlation between inventory forecasting and revenue forecasting.) We had to get it as right and as real time as possible. With a new forecasting strategy I was able to predict the size ratio demands for the different categories of inventory using exponential smoothing (essentially an algorithm of regression with seasonality adjustments) for styles with a longer data set and weighted rolling averages for newer styles. I took that further in future iterations, by comparing a new style that had an affinity with older styles and adding an adjustment into the rolling average weight, and I extended the forecasting process to include feedback from field sales for making macro level adjustments on demand. At each stage I would do post-hoc testing to validate the algorithm changes using experimentation against known results. The net result was significant reduction in returned bras, and a big increase in sales because prior to the project we were selling out of the correct sizes.

So how does this relate to inventory optimization in the online environment? Everything...think about the correlations; forecasting at higher grains, field sales integration, data mining, algorithmic strategy and modeling, post-hoc testing, simulation, ephemeral inventory, seasonality, inventory performance affinities, margin hits on channel routing to remnant etc...

Like 1998 in manufacturing the online industry is rapidly innovating on inventory (read "yield") optimization. Sophisticated solutions are popping up along with some rising startups; The Rubicon Project, Yieldex, Pubmatic, AdMeld to name a few. Each with a different flavor for improving yield, some more technically oriented, some more strategically oriented. Throw into the mix multiple channels (and channel conflicts, oh boy fun) for selling the inventory; in-house sales, ad networks, ad exchanges, of which I know a few things ;), and you have an amazing transformation in the industry. Now we just have to figure out how to make it all work together. So let's get back those forecasting books. Here is a refresher on some key concepts:

Forecasting Methods
  • Extrapolative Methods - Find a pattern(trend) to the historical time series and assume that the pattern will continue into the future.
  • Explanatory Methods - Determine the factors which explain the past behavior of the variables to be forecast.
  • Judgmental Methods - Methods which rely, not upon a statistical analysis of the historical data, but on Managerial expertise, feedback from Sales Reps and Customer surveys.

My experience is that you use all three and over time strive to find the right mix of the three and the right mix of possibilities within the three.

Forecasting Indicators

  • Leading Indicators - Provide advance notice
  • Concurrent Indicators - Provide confirmation at the time of the event
  • Lagging Indicators - Provide notice after the event

While it is tempting to dismiss Lagging Indicators, they are useful for post-hoc tests of new econometric models (Extrapolative Methods) due the "Assumption of Constancy" (the patterns of the past indicate the patterns of the future.)

Special thanks to Dr. Len Tashman director of the Institute for Forecasting Education and editor of Foresight - The International Journal of Applied Forecasting for his many years of support and mentoring. Contact the Institute if your company is looking for an evaluation of your revenue and/or inventory forecasting methods.

Fair Use Policy: Please include the name: Edward Boettcher and the following link: The Product Messenger when referencing this information. All content (c) 2009 Edward Boettcher

Monday, August 10, 2009

Learning to Love Opt-In

As the story goes there has been a mad scramble in online advertising as of late to mitigate for pending changes in the advertising ecosystem. The "wild west" of behavioral metrics tracked in the cookie with limited regulation has raised the brow of the FCC. The IAB and other trade groups are stepping up efforts to thwart government intervention via stronger self-regulation efforts. The doomsday scenario being a government mandated, strict opt-in that has the potential to bring down big players in the ad network game, particularly those who have invested millions in behavioral targeting technology, as well as crush the bottom line of publishers.

In other news...Rupert Murdoch, faced with disastrous revenue projections from his online, newspaper and magazine media properties has decided to stop the bleeding and charge for online content. Other news and content providers who have seen their content shared and shared again as their product is propagated throughout the Internet providing monetization that doesn't seem to end up back in their pockets will no doubt quickly follow suit and the Internet will turn into a pay to play entity for the information consumer. Free content as we know it will all but disappear.


The Internet and the advertising ecosystem will evolve. Embrace it. Learn to love opt-in. Opt-in provides freedom, freedom brings opportunity and opportunity will shape your future (free prize to the first person who e-mails me what movie that line comes from.) Allow me to explain. For years online media companies have been upping the ante on collecting data on Internet users in a very opaque manner. While use and control of the information has been generally well shepherded by the companies involved, greed will always lead to an escalation that pushes to a tipping point somewhere and that tipping point happened with NebuAd, Phorm and others when they planned on and/or started looting ISP's for user information without user consent. Our trust as consumers has been breached and there is no going you what...tell me what's in it for me, maybe we can work some thing about I give you my personal data and purchasing intentions up to a level I'm comfortable with and you give me content and other goodies in return. Quid pro quo. What do you say?

As an agency, data provider, marketer, media buyer, ad network etc... I say yes! Hallelujah! We are about to witness the greatest cooperation between the buyer and the marketer in the history of advertising. "Free" content will soon turn into a quadrangulated transaction between the marketer, the consumer, the data enabler and the content provider.

Allow me to introduce some new terms:

Lightning - The connection between the cloud and the earth. The combined data set of online derived user purchasing behavior and brick and mortar and other offline derived user purchasing behavior. Lightning is powerful and a "tall" consumer will install a lightning rod via opt-in. Think coupons downloaded onto a smart phone and redeemed at the store followed by a thank you e-mail with additional purchase incentives in a virtuous cycle.

Opt-in Leveling - Systematic collection of increasingly higher grain and more verifiable user supplied behavioral information. Today, give or take how you define things, we can get up to 3 levels of behavioral information. Level 1 would be a minimally populated cookie jar. Level 2 would be a richly populated cookie jar containing your cookie and perhaps other cookies that can be linked to a 3rd party data provider. Level 3 would be Level 2 plus a purchase or registration with your site that you could possibly append to PII that could be used to enrich future conversations with the user. More on this in a few milliseconds...

User Driven Targeting & Optimization - User selected indications of segment, purchasing behavior, brand affiliations and other behavioral metrics. I envision the rise of opt-in leveling tools that feed targeting, re-targeting and optimization.

In today's environment we can get to a Level 3 but there it starts hitting the area where the consumers trust is being breached. With Opt-in we can get to Level 10 and beyond. Conversations with the consumer can be rich and rewarding for both the marketers and consumers and profitable for the intermediaries that help make the final transaction happen while supporting continued access to quality content. Look at the trends in social networking, particularly Facebook and MySpace. People are not inherently averse to sharing personal data, but there is a comfort zone that could possibly be negotiated. If I notify an intention to buy a home in an opt-in leveling tool then I shouldn't have to pay for content, that fee should be provided by the mortgage brokers. And if I'm giving up additional PII, I want a piece of that lead gen fee, not some vague promise of a free iPod. Quid pro quo, enabled by opt-in.

Fair Use Policy: Please include the name: Edward Boettcher and the following link: The Product Messenger when referencing this information. All content (c) 2009 Edward Boettcher

Thursday, April 23, 2009

Bringing Our Life to Work

Welcome to the first post on The Product Messenger. If you are a Product Manager or involved with Online Media, I hope you find this to be an entertaining and valueable site for discussion.

I’m sure most of us have found ourselves at home eating dinner, or at the gym etc… still ruminating about an issue we had tackled at work that day. In a dynamic industry such as Online Media, the distinction between the working world and our personal world is often blurred. For me personally, I gave up caring about that distinction when I started my own company where my working life and personal life were perfectly entwined. Now that I’m back as an FTE in the corporate world there is a little more separation, but I find that some of my best ideas happen outside of work while I’m hiking or doing whatever.

Now to the point. While I think many of us ponder the effect work has on our personal life, I think it’s a good exercise to identify things in our personal life that strengthen our working life. I find great value when interviewing a candidate or interacting with a collegue to find out what they like to do outside of work.

When I am home and not spending time with my Wife, you will generally find me in my recording studio playing and/or recording and/or mixing audio. If I’m not in the studio I’m usually building or restoring audio gear for the studio. I bring this up because those activities very much inform and enrich my outlook as a Product Manager. Building a studio and making an album sound great is an incredibly complicated process. A great deal of visualization happens and the realization can often take months or years to support your initial hypothesis. Below are two simplified diagrams that show just how close the thought processes can be. Diagram 1. outlines a typical data extract, transform & load (ETL) process between some action and a downstream data store. In Diagram 2. you see a typical signal path during audio recording.

My question for discussion is: What things do you do outside of work that make you better at what you do inside of work? Please comment.

Diagram 1.
(Click image for larger view. Back button to return)

Diagram 2.

(Click image for larger view. Back button to return)

Fair Use Policy: Please include the name: Edward Boettcher and the following link: The Product Messenger when referencing this information. All content (c) 2009 Edward Boettcher