Jonathan Mendez's Blog

The Relevance Delta: Madison Avenue and Google

Andy Berndt’s departure from O&M set the tone for Advertising Week. If you were present at MIXX and OMMA and wanted to find out how advanced advertising has become you were in the wrong place. I direct you Mountain View, California where they are eating Madison Avenue's lunch. What I saw the last few days was a troubling view of the future balance of advertising power.

Case in point is Digitas. With eager expectation I listened to Digitas speak about their work for IBM (full disclosure - IBM is also a client of OTTO Digital). What is Digitas doing for IBM? For one they are adding links to social media sites like Digg and del.icio.us to each piece of content. This they said creates a viral effect. The other strategic nugget was that they were able to deliver relevant content AFTER the user registered on the site.

Digitas and IBM are smart organizations. They are “digital.” They are supposed to get it. What about targeting IBM’s content based on source data? What about targeting IBM’s content based on visit behavior? Why does IBM need 30 websites and 15 CMS systems when they can have a single platform and geo-target everything? Why not test before deciding each country having their own creative and messaging was a bad idea? Why not think, deliver and create solutions like a search engine?

What about the consumer brands? Here, video was on the minds of everyone likely due to the fact that brand dollars will roll into digital through video. So, how is Coca-Cola using video in interesting ways? Coke created a “made for the web” video that was so well liked that they used it for a TV commercial. Where is the value creation here outside of reduced creative costs? Where is the value creation with their customers?

- Links to Digg
- Targeting Content to Registered Users
- Putting Online Video on TV

Digital marketers need to do better. Digital marketers need to create value. I don’t care if you are a brand marketer of direct marketer. Value is created by delivering relevance. Google is laser focused on this. After two days at MIXX and OMMA what I can tell you is that this relevance delta is getting wider and wider.

The solution for digital advertising, digital marketing and digital publishing is in products. Products solve problems. Build solutions – be it with your own technology, the technology of others or just leveraging existing platforms. This stands not only for agencies but also for the businesses that employ them.

Take YouTube as an example. I can find just about anything I’m looking for on YouTube. Flinstones, T-Rex or Queens of the Stone Age - YouTube is a video product that delivers relevance. YouTube is a platform. YouTube is a widget. YouTube is everything that Madison Avenue is not creating.

At the end of the day it is this marriage of technology with great creative that will deliver relevance. Why does this work? Look how Google approaches advertising.

1. Viral & WOM- They have never spent a penny on an ad yet have one of the most valuable brands in the world
2. Insane focus on the user – I didn't hear the words “user experience” uttered once in two days. Note that 1 & 2 are inseparable.
3. Optimization – Publisher side, ad side-- create then test, measure and optimize the hell out of it.
4. Technology is the core – Success begins and ends with the platform

Andrew Chen has a great post on the cultural differences between Silicon Valley and Madison Avenue that he saw the past two days at MIXX. The issues that divide the coasts on the product/solutions front are likely greater. Maybe this is a byproduct of the different mindsets however, increasingly these worlds are not coming together.

While Madison Avenue has most of the dollars in their hands now, the purchase of aQuantive by MSFT and the rise of in-house SEM over the last few years due to Google’s interface, reporting and support are just two events that should motivate the agencies to create marketing solutions in new and better ways. Measurable value creation and ROI for clients is achieved by delivering relevance. Relevance is only delivered through technology.

How many products and marketing solutions are you building next year? How many is Google working on? How wide is the delta?

Advertising Week 2007

Next week I’ll be at what is referred to as “Advertising Week” here in New York City. On the Digital side OMMA and MIXX have their shows scheduled right against one another. Last year I spoke and attended OMMA. I was a bit disappointed with the show however my presentation from it; Optimizing Content Pages - When Every Page is a Landing Page is by far the most viewed presentation I’ve posted on this blog.

This year I’ll be moderating a session at MIXX and attending both shows. John Broady, our newest Client Partner at OTTO Digital who comes to us after being VP of Operations and Business Intelligence at CNET will be speaking at OMMA along with Matt Roche, CEO of Offermatica. If you’re going here’s where we’ll be:

OMMA: Monday 4:45pm
Follow the User: Leveraging On-Site Behaviors
Every visitor and every page she views at a site offers publishers a clue about how to better serve her, and yet very few sites even alter their home page for new and returning customers. Most of our sites literally sit on piles of behavioral data that could be leveraged to serve more personalized content to a user, focus advertising strategies or just make a visitor feel more welcome. How are the most advanced sites realizing real revenue gains and better customer relationships by tracking user behavior? And how are they taking raw data and turning it into actionable knowledge about customer?

Panelists:
Matt Roche, CEO Offermatica
John Broady, Client Partner OTTO Digital
Jack Jia, CEO Baynote
Jeff Hirsch, Chief Revenue Officer, Revenue Science

Moderator: Steve Smith

MIXX: Tuesday 1:45pm (Room 402, 4th floor)
The Next Generation of Applications for Targeting, Personalization, and Advertising
The pages and paths approach is giving way to a world where the delivery of relevance to drive engagement rules. Nowhere is this more evident than in the rise of widgets and applications. Next generation widgets and applications are already being developed that use quantitative data to optimize and target the delivery of contextually relevant content across experiences. In Social Media, this creates a viral effect. For publishers, the rise in engagement metrics enables a more detailed discovery of user behaviors resulting in deeper more valuable segments. For advertisers, this means better targeting. This workshop will focus on the applications, strategies, measurement and benefits that can be expected in this emerging medium that looks to be transforming the web as we know it.

Panelists:
Sam Bloom; General Manager, Interactive, Camelot Communications
Andrew Chen; Entrepreneur-in-Residence, Mohr Davidow Ventures
Andy Monfried; President & Founder, Lotame
Kevin Tate; Founder and Principal, StepChange Group

Moderator: Jonathan Mendez, Chief Strategy Officer, OTTO Digital

OMMA: Tuesday 2:15pm (Gramercy, Suite A)
Crunching Numbers - The Growing Importance of Data-driven Decision Making
Gone are the days of intuition. Welcome to an age where data is changing the rules of the game. Today's best and brightest marketers have turned into savvy statistical practitioners, harnessing the power of data to shape their decisions and help them achieve their business goals. Author of the new book Super Crunchers and Forbes columnist Ian Ayres will discuss why thinking by numbers is the new way to be smart. He will be joined by The New York Times and eHarmony who will share real-life experiences detailing how data shapes their organizations and helps optimize their businesses. You will not want to miss what promises to be a lively discussion.

Panelists:
Ian Ayres, author of Super Crunchers and Yale Law School professor
Janeen Wasoski, Advertising Director, Digital Sales Strategy, The New York Times
Jon Ward, VP, Business Development, eHarmony

Moderator: Chris Duskin, Director Product Marketing, Offermatica

Landing Page Resolution: 800x600 vs. 1024x768

While not one of the eternal questions of our time, what screen resolution to build your landing page is nevertheless always a key design consideration. Many factors weigh into this decision. What are the browser resolution settings for the majority of your users? How does the page appear on different resolution setting (e.g. where is the call to action)? Or the worst one of all, what does the design or style guide permit?

Recently, we worked with a client whose existing landing page designed at 800x600 was performing very well. I had the idea that building multiple landing pages optimized for different browser resolutions and then targeting the pages to match the user’s resolution might improve conversion rates further. However, before we went about creating this layer of personalization we needed to test and validate this assumption. In the interest of time and resources we built one new page in 1024x768 (rather than three that would include the higher resolutions). We created performance segments for each of the four popular screen resolutions and put the pages to the test. The results are quite interesting.

At the lower resolutions the results are counterintuitive to my targeting hypothesis. Visitors with 800x600 actually converted at a significantly higher rate on the 1024x768 page. Those with 1024x768 resolution performed better on a page that was built in 800x600. Hmmm…

For the higher resolutions the effect was more expected. 1280x800 visitors converted at an almost 2% improvement on the 1024 page. Moving up to 1440x900 the improvement factor for conversion rate was over 6%.

What does this tell us?

800x600 Visitors: Since both versions of the landing page were designed to have the headline, fields and call to action visible without scrolling in 800x600 we can hypothesize that the larger presentation on these key landing page factors on screen accounted for the over 5% lift in conversion for the 1024 page on visitors.

1024x768 Visitors: This outlier is frustratingly difficult to understand. One hypothesis is that users on 1024 are much more likely to keep their browser windows not at full display mode. By shrinking the browser window it’s possible that the 8x6 page provides a better experience since no scrolling would be needed to interact with the page. A follow up test with some custom JavaScript will let us know if this is indeed the case.

1280x800, 1440x900 Visitors: Our original hypothesis remains strong in the higher resolutions. The larger resolution on landing is clearly has an increasing factor as you move to these levels. A follow up test would be creating pages optimized for these higher resolutions. There were many visits from laptops and work environments with these resolutions. Understanding the factors of higher resolution browsers more fully will be very beneficial.

Does it make sense to target the landing pages?

Sure does. Targeting these landings should yield a compounded conversion rate increase around 13% based on our initial testing. Additionally we’ve validated a fundamental design change for our client and continued testing as mentioned above will likely yield continued improvement.

What should I do?

Create pages in different resolutions and test! Screen resolution is a fantastic high impact segment. Nice lifts in conversion and other performance metrics can be easily gained by targeting the most basic (and most important) user experience variable, screen resolution.

Why Mashups are Mandatory for Marketers

No longer is the web about sites. It has become a database platform with lots of structured information. This change is profound. The strategic and creative use of this structured data (from REST APIs) is changing the way we interact with the web… and for the better. Deeper and more engaged user interactions are taking place and breeding higher levels of emotional connection to digital (hello brand marketers) because as the web becomes more intelligent it becomes more relevant.

These experiences, call them applications or mashups (and their little sister, the widget) are unfolding around us. The higher degree of relevance is helping businesses succeed in the space where your competition is only a click away. How can your marketing begin to benefit? It begins with the understanding that with the rise in APIs and data streams no longer are we optimizing each page to accomplish a single user goal. Mashups provide the ability for primary, secondary and latent user goals to be achieved simultaneously. This is what fuels deeper engagement and raises performance metrics.

The most prominent and clearest example of this change is the one two judo chop of Google’s Personalization and Universal Search. No longer is the success metric for Google getting users off the page as soon as possible. Now Google would like to keep them on the page engaging them in a number of different content sources and dynamic experiences. The ultimate goal, delivery of relevance to user intent has not changed. However, Google has learned that the ability to deliver relevance in a more holistic dynamic experience rather than a multi-page flow improves the user experience and thus their business results. When a company like Google that was built on user experience and testing makes this type of change it marks a titanic shift.

So Why Do You Need to Do a Mashup?

1) Mashups are intrinsically strategic. We first need to know what structured data is at our disposal and then we can strategize about the best ways to design and deliver it around the user goals to create a seamless delivery of relevance.

2) Mashups are inherently social. They take pieces of data from disparate sources and put them together in useful, problem solving ways. This socialization of content works on two distinct levels with mashups. First it allows an organic and viral growth to occur for the application (the idea behind widgets). Second, it allows these applications to benefit from the collected wisdom of the community, the individual contributions of members or both (the idea behind Universal Search).

3) Mashups are deeply engaging. They provide a richer more dynamic experience. Mashups keep users engaged or “in the flow” of whatever goal they are trying to achieve by providing a greater ability to deliver contextual relevance. Mashups also drive relevance through user controlled intention refinement that allows the targeted delivery of content.

4) Mashups are personal and targeted. The ability to provide multiple dynamic refinement of content and optimize based on real-time self-segmentation delivers experiences that are more closely aligned with delivering targeted content in real-time based on what we are learning about user source and behavior.

5) Mashups perform better. Our initial work on application and widgets this past year has yielded huge lifts in conversion and engagement metrics. These include a widget to optimize content distribution from Social Media, a search interface mashup for one of the major engines and a driving directions mashup for a leading digital mapping provider. Increases of 150-400% in key metrics are what we’ve experienced which is probably why we’re working on more of these for clients with no end in sight.

6) Mashups are quant. The amount of data collected, testing and optimization that can be done on mashups is astounding and the changes to implement ideas are much easier and can potentially have a greater overall impact. The dynamic nature actually lends itself to testing better than static environments.

7) Mashups are the future of SEO. So says Sep Kamvar head of Google’s personalization team on the panel I shared with him at SES San Jose and personalization is where Google is going with search. The mashup delivers greater relevance to users and an improved user experience. Both key factors for Google. Also mashups and applications can be easily widgetized and ported. As web authoring tools allow more client side customization the value of applications and widgets will rise and they will weigh more heavily into Google’s personalization algo.

8) Mashups are the future of SEM. The landing page as we know it is dead. The amount of relevant source data from paid search that can be used to deliver relevance dynamically through APIs is astounding. We most also take into account that personalization algos will soon creep into the serving of search ads and display and univeral search will soon allow rich ads - more of a reason to create mashup landings.

So to paraphrase George Clinton, “Free your data and your results will follow.” Once an API frees the content it becomes infinitely more powerful because it can be used much more strategically. Your client doesn’t have an API yet? Make one!

Digital marketing is about the marriage of content and technology. The cornerstone of this is taking data and delivering it in new and interesting ways to deliver relevance. If you’re not using the platform to do this you’ll soon be under it.

***More Thoughts: I will be speaking at two upcoming conferences on Applications, Mashups and Widgets. On September 25th in NYC at MIXX where I will be moderating the panel The Next Generation of Applications for Targeting, Personalization, and Advertising.

On October 6th I’m excited to be in Silicon Valley at the CommunityNext Platform Conference where I will be speaking at the session Metrics and Targeting.

***Must Read: Andrew Chen of MDV who was nice enough to invite me to speak at the upcoming CommunityNext conference has a brilliant post on viral marketing strategy.

Optimization of Ads & Landing Pages by Keyword Match Type

Exact Match, Broad Match, Phrase Match – Understanding Relevance and Unlocking Higher ROAS

With the advent of quality scoring for ads understanding of match type has never been more important. I’m not referring to understanding how match type works from an ad serving perspective but understanding how match type drives clicks with different goals and at different stages of the consideration or purchase funnels. This impacts both your ad performance and landing page optimization strategies. So, if there is one factor of your search program you need to start understanding to improve ROI it is match type.

Recently at OTTO Digital we did a landing page test from paid search ads and created profile parameters that provided segmentation of the results based on exact match, broad match and phrase match. The findings were incredibly interesting and speak volumes about the needs of relevance (and thus conversion rates) based on match type.

The optimization we did was a simple reinforcement of the query. It looked like this on the landing page.

We’ve done many tests on reinforcements of the query (it is parsed from the referring URL and generated on the page using Offermatica) and the findings have validated my theories of the need to bridge the moment of recognition with relevance for the user. As you can see from the results filtered by match type, the less relevant the query was to the ad, the more effective keyword reinforcement was.

Results below are measured against the existing landing page that had no reinforcement:

• Exact Match +2.63% Conversion Rate

• Broad Match +4.36% Conversion Rate

• Phrase Match +15.16% Conversion Rate

While the benefits of keyword reinforcement were clear for exact match, the performance increase was even more striking for broad match and phrase match; the places where relevancy is intrinsically more challenging both ad side and on the landing page.

I hypothesize that the huge lift in conversion rate for phrase match seems to be related to the dichotomy of more specific queries but less relevant ads. Users from phrase match are hoping their query string will deliver a higher degree of relevance in the results. However, match type is rendering ads that are more generic in nature. Seeing the relevance of their query delivered on the landing page clearly had a large positive impact for these users.

The point I’d like to make is not that everyone should start doing query reinforcement. Due to its simplicity (and results) I'm using this strategy to help clearly illustrate a larger point. The relevance of your messaging and creative both ad side and on the landing page will vary for the same keyword based on match type. With the advent of quality score forever changing ad serving and pricing models in all three major search engines looking at performance data by match type and testing ads and landing pages by match type becomes essential. Strategies here to deliver relevance and improve your return on ad spend (ROAS) will only continue to become more amplified.

One easy way to start testing both your ads and landing pages is to create AdGroups separating the keywords by match type instead of bundling them together in a single AdGroup as is done most often. Distinct parameters can also be added to the referring URL to specify match type. These are the first steps. There are many other ways that we are testing with our clients to optimize ads and landing pages by match type and in the coming months I hope to share more with you. In the meantime, start testing because if your competitors are not already they will be soon.

Note: While I brushed off understanding how match type works in the engines (and it works a little differently in each) it is incredibly important. Here is a great post by Chris Zaharias from a couple of months ago on recent Google match type changes and how it impacts their bottom line.