Don’t Just Think Digitally, Act Digitally

November 6th, 2009 View Comments

Note: This post was originally published in the AdAgents column of AdExchanger.com. AdAgents is an opinion column written by agency-side members of the online display advertising community.

Digital media has caused huge shifts in the advertising industry, including the rise of the Digital-Driven Media Agency — media shops that have identified digital media as the key strategic focus for their organization. Management teams at these agencies have set aggressive goals to increase the share of billings coming from digital. Many agencies have collapsed separate print, broadcast, and online divisions, instead forming integrated planning groups led by “digital thinking.”

Agencies have recognized that in a landscape where consumers control their own media consumption, brands need to be authentic, interesting, and useful. In response, they’ve formed branded entertainment and social media teams to help brands thrive in this new landscape.

Now it is time for agencies to go beyond thinking digitally and begin acting digitally. Agencies have embraced digital in terms of high-level corporate strategy but technology isn’t woven into the operational fabric. Planners think digitally, but often only have dated tools like phone and fax available for actual execution. As a result, it is an outside vendor who is actually integrating the data feed for the dynamic ad, determining the optimal frequency and optimizing the targeting parameters. Of course, it’s in neither the agency’s nor the client’s best interests to have the agency do everything. But to act as effective stewards of the client’s interests, and to make a compelling profit while doing so, agencies need to bring more technology within the walls of their own organizations.

Fragmentation, for example, is a defining trait of digital media, and agencies have not adopted significant technology to buy this fragmented channel effectively. Purchasing media over phone, fax, and e-mail introduces a major constraint on how many sites can be included on a media plan. Agencies themselves can’t aggregate a digital audience that is both sufficiently large and sufficiently targeted so they turn to ad networks. Once ad networks are responsible for a portion of spend they are also responsible for the optimization of that spend, which they are well-equipped to do.

Agencies have handed two key business opportunities around digital media: aggregating audiences and optimizing spend based on response to outside partners with the required technology. The most important issue here isn’t even that agencies are passing over two very high-value activities at a time when they are struggling to achieve acceptable profit margins. What is most problematic is that audiences are being aggregated and campaigns are being optimized based on sell-side interests, instead of aligned client and agency interests. Mike Nolet of Appnexus does a fantastic job of explaining the implications of this flawed incentive structure. (See the post.)

To be clear, I am not suggesting that agencies become technology companies or masquerade as such. Instead, I am saying that agencies have not adopted sufficient technology to continue in their historic role of reasonably compensated agents of the client’s interest. It is a positive indicator that forward-thinking agencies are spinning out separate business units focused on data-driven targeting, just as aQuantive did with DrivePM in 2004. But the real victory isn’t creating island of excellence where a select few are able to leverage technology, sometimes proprietary, on behalf of clients. The real victory, and the greater challenge, will be deeply integrating technology into the day to day activities of everyone at the agency who is thinking digitally.

Foursquare as a Mobile, Social, Content Platform

August 18th, 2009 View Comments

I’m a big fan of mobile social network FourSquare. It’s a very versatile service, but here’s a quick list of what FourSquare offers its users:

1) A mobile social network that allows you to broadcast your location to your friends and see where your friends are. You can check in from anywhere you want on FourSquare but public places are most popular.
2) A game that rewards you with points and badges for checking in at bars and restaurants, and lets you compete with friends
3) A personal record of all the bars and restaurants that you’ve visited (hat tip: Charlie O’Donnell)
4) A directory of all the bars and restaurants near your current location
5) A social recommendation service for restaurants, bars, and other cool spots to check out.

The recommendation service is what’s most valuable to me. When I open up FourSquare it automatically determines my location. In addition to seeing a list of nearby bars and restaurants, I can view specific “tips” and “to-do’s” — think: recommendations — for places nearby.

For example, if you get out of the subway at the Christopher Street in the West Village and open up FourSquare, here are some of the tips that you’ll see:

foursquare screenshot

Sure, its possible to visit Yelp.com on your computer and find the best restaurants in a neighborhood before you go there. But when you find yourself hungry or thirsty in a random neighborhood, its great to have these recommendations on demand on the phone. Being able to filter the FourSquare community’s tips so that I only see my friend’s tips is a great feature. When I’m deciding where to go, nothing could be more relevant than finding out that one of my friend’s favorite restaurants is a few hundred feet away that there is one dish that is better than everything else on the menu. FourSquare delivers this relevance with only a few clicks.

The quality of the tips generated by the Foursquare community is very good, but ultimately it’s the content delivery method — mobile, location-aware, sorted by distance and social relevance — that’s most valuable to me. Being able to pull in tips from favorite external sources, like Thrillist, Daily Candy, Chowhound, Serious Eats, Urban Daddy, would make FourSquare even better. I would love to see FourSquare act as a platform where I can subscribe to specific review publishers and access their content (free or paid) into the Foursquare interface.

Most tips are as brief as a tweet, which is a much better length for mobile consumption than your average multi-paragraph restaurant review. But I could easily imagine an ecosystem where it is profitable for publishers to reformat their content for consumption in the FourSquare interface. Without any hesitation, I would pay a $20 one time fee to have the full archive of Thrillist reviews accessible in FourSquare. I have hundreds and hundreds of historic Thrillist e-mails archived in my Gmail that provide zero value to me and probably aren’t providing much revenue for Thrillist, assuming that most of their ad revenue comes through daily e-mail blasts as opposed to the historic content on their website. Reformatting reviews for mobile consumption would unlock a lot of value for both me and Thrillist.

I would also love the ability to pull in the restaurants that my friends and I tag on Delicious. This is rich information that would be much more valuable if I could easily accesss it in the FourSquare format: mobile, location-aware, sorted by distance and social relevance.

Optimizing Beyond Click and Conversion

July 14th, 2009 View Comments

In my first post, I wrote about how advertising optimization allows marketers to increase ROI by shifting allocations between audiences, media types, and creative versions. Usually these three areas are optimized against direct response metrics: the click and the conversion. But the use of automated optimization should not be synonymous with a myopic focus on click and order counts. It is possible to optimize campaigns to do more than generate clicks and online sales. If you can measure the persuasive, memorable, emotional impact of an ad — impact which clickthrough rates and conversion rates alone does not capture — you can optimize campaigns to better build brands.

But first, let me give the click it’s due respect. For one thing it’s immediately measurable. Brand study survey results might not be available for weeks, but the click is measured instantaneously. Second, attribution is clear. The adserver knows which exactly which site delivered the click, so its straightforward to optimize to deliver more clicks. Lastly, there’s no selection bias since every click is captured.

Its difficult to find a branding metric as easy to incorporate into optimization models as a click, but here are some other metrics that agencies can use to optimize branding efforts.

1) Ad Exposure Time — One important difference between ad impressions is ad exposure time. Lotame has published a white paper, “Time Exposure by Banner Size,” that found that 300×250 ads were visible for an average of 13 seconds, compared to an average of 5 seconds for 728×90’s and 2 seconds for 160×600. In addition to creative size, ad exposure time varies greatly by site type. Banners on blogs and social network pages will have much lower exposure time than a companion banner on a video site. VideoEgg recently released their floating Twig ad unit in part to address the challenges that blog publishers face given the low ad exposure time.

All else being equal, an ad that is visible to the user for 20 seconds should have a larger impact on brand health than an ad that is visible for 2 seconds. This type of information should be incorporated into advertisers’ impression bidding strategies.

If a planner is building a branding campaign, they should be able to target sites in their planning process where the ads are visible for a longer period of time. At the very least, they should be able to ensure that if their advertiser’s name and logo only shows up in the 7th second of animation, that each impression is visible for least 7 seconds on average. In a buying platform, planners should be able to enter in time exposure rules by creative. There also needs to be a comparison between value (average ad exposure time) and price (CPM). This is where planners need to rely on the automated optimization power of a buying platform.

2) Time Spent on Landing Pages — Not all clicks are created equal but just as click traffic for DR campaigns can be differentiated by conversion rates, clicks for branding campaigns can be differentiated by time spent on the site. Optimizing audience targeting by clickthrough rate is a risky strategy when studies have shown that a small majority of users generate the majority of the clicks and that display exposure still has a significant impact on those who don’t click. For branding campaigns, time spent on the landing page is a much more meaningful metric for understanding which audience segments are interested in engaging with the brand. By incorporating time spent metric into automated buying platform, an advertiser can pursue a pay-for-engagement buying model on any type of media with any type of creative assets.

3) Visits to Advertiser Homepage — A consumer who doesn’t click on an ad may still visit the advertiser’s homepage to learn more about products. Measuring the number of impressions it takes to generate a visit to the homepage is good measure of ad effectiveness. This metric is especially important for multi-channel retailers where online advertising drives consumers to do further online research but end up buying in-store. The impact of online advertising on these particular consumers is never captured by online click and conversion metrics.

There will never be an algorithm to spit out innovative, memorable integrated marketing programs, but marketers still have the opportunity to use automation, optimization, and data to build brands.

People, Not Pages? We Need Some New Measurement Tools

June 9th, 2009 View Comments

This week comScore announced the release of comScore 360. Instead of estimating audience size solely off a 2mm panel, comScore 360 will operate on a hybrid model — information from partner sites that implement the comScore beacon will be layered on top of panel data. This should give media planners a better view of audience size compared to just site user counts, which are inflated by cookie deletion, or just panel information, which is distorted by the relatively small sample. For a better understanding of how difficult it is to measure online audience, I recommend Quantcast’s Cookie Corrected Audience Data whitepaper.

A hybrid model is a big step forward for the “Gold Standard” of audience measurement tools, but it still doesn’t address the new challenges faced by media planners. As the “People, not Pages” approach gains greater acceptance, media planners will be less focused on choosing one media property over another. Instead of choosing between Maxim.com and AskMen.com, planners will choose between two ad networks offering the same segment targeting men 18-25 with an interest in fashion. But what is the reach and composition for each of the two ad networks offering that same segment? Right now, there is no independent 3rd party audience verification tool for targeted media. This is fine for direct response marketers, where targeting is simply a means to a very measurable end: online response rates. But for brand marketers who define success as reach and frequency against a key demographic, independent audience verification is key. Focusing on “People, Not Pages” allows marketers to be much more accurate and efficient, but marketers need assurance that they are getting what they pay for.

How do you see this issue being resolved? Through survey companies like Vizu and InsightExpress? Measurement companies like comScore? Pure-play data vendors like BlueKai? Data/media vendors like Aperture? I’m curious to hear your thoughts.

The Three Pillars of Digital Advertising Optimization

May 31st, 2009 View Comments

When someone mentions “digital media optimization”, what do you think of?  An A/B creative test? The advanced algorithms that make one ad network better than another?  Pulling a report halfway through a campaign, calling publisher partners, and sending faxes to reallocate to better performing sites/placements?

Almost all digital media campaigns involve some element of optimization, and in the next 2-3 years optimization will become more pervasive, more rapid, and more automated.  Digital media planners and account managers will still have control over their campaigns, but they’ll spend more time considering strategic issues and less time calculating and comparing the ROI of different media options.

It’s hard to say who’s going to be the leader in the optimization space: media holding company ventures like Havas’ AdNetik, Vivaki Nerve Center, and Varick Media Management; dynamic creative services like Tumri, Adisn, and SnapAds; ad networks like Turn, Tribal Fusion, and X+1; or buy-side solutions like InviteMedia and Media Math.  The list goes on and on.  There are a lot of players who increase advertising performance by optimizing better than the next guy.  But when I think about Optimization, here are the three pillars that I use to frame the issue:

1. Audience

2. Media

3. Creative

Audience:

This the individual consumer that you are reaching with the ad.  These are the traits that would be relevant to optimization:

A) Demographic profile (age, gender, DMA, HHI, education, etc.)

B) Data from previous ad interactions (first impression vs. third impression, previous visits to campaign landing page, previous visits to advertiser homepage)

C) Data from transactions with brand (current customer vs. acquisition prospect, customer data from client CRM database)

Media:

This is the environment in which you are presenting the ad.  These are the traits that would be relevant to optimization:

A) Media property (Major media brand vs. long-tail site, blog vs. social network vs. editorial site)

B) Page placement (Above the fold vs. below the fold)

C) Context (If a semantic engine scanned the page, how would it classify the content?)

3. Creative

A) Format (Standard flash vs. rich media vs. video)

B) Creative Message (What product are you promoting?)

C) Creative Details (Color of ad, call to action, intro animation, logo placement)

In my experience, media optimization is the most common but there are huge gains to be had by optimizing all three, and even greater gains for those who optimize all three in an integrated platform.

Bottom Line: More and more of digital media will be optimized in the future. Marketers should look at the three pillars of digital advertising optimization — audience, media, and creative — and ask themselves: Which pillars am I currently optimizing? Which pillar do I understand best when it comes to controlling performance? Which pillar would benefit the most from quicker, more automated optimization?

Newer Entries »