Efficient Frontier Insights

SES London!

Search Engine Strategies returned to London last week, but in a new location – the Queen Elizabeth II Conference Centre – just a stone’s throw from iconic attractions such as Westminster Abbey and the Houses of Parliament. But was the content inside as impressive as the views outside?

I spoke on two panels.  The first, on Tuesday, was called ‘PPC or SEO? The Ultimate Search Marketing Battle’ and was a fun, knockabout session to pit the worth of PPC against SEO.  I waved the flag for PPC while SEO was ably defended by Dave Naylor of Bronco.  Bill Hunt of Back Azimuth Consulting gave insights on the interaction between the SEO and PPC while the session was moderated by Jon Myers of Yahoo!

Beyond the verbal sparring between Dave and I, what is clear of course is the value of understanding the interactions between different media channels, including SEO and PPC.  A continued focus for Efficient Frontier in 2011 will be the development of tracking and optimization algorithms to help our customers understand and maximize these critical cross-channel interactions.

The second panel, on Wednesday, was called ‘Ads in a Quality Score World’.  Chaired by Brad Geddes of CertifiedKnowledge, we took a deep dive into the Quality Score and best practice to improve it.  Paul Mead of VCCP Search gave an excellent introduction into the theory and practice of how Quality Score works together with case studies. 

Then I presented on how to deliver great Quality Scores at scale across large, enterprise search campaigns.  Crafting each adgroup by hand is not an option when you have 150,000 adgroups and 1.3 million keywords, which was the example I used from one of our UK travel advertisers.  The key is to use best practice with the right structures and templates, from which high quality campaigns can be built quickly out using advanced management tools such Efficient Frontier’s.

Overall, there seemed to be a new buzz about SES.  Whether that was the new location or a renewed sense of optimism this year was unclear.  But for advertisers, practitioners and technologists from all over Europe, this clearly was a valuable event to attend.  I’m looking forward to the next one!

LET'S TALK DISPLAY STRATEGY

Over the past year or two, many advertisers who rely heavily on digital marketing as part of their direct response strategy that has traditionally focused on search engine marketing (SEM) and Email advertising have been dipping their toe into Display advertising.  Historically, those advertisers have viewed Display as part a branding channel of the digital ecosystem based on the silo approach of direct response and brand advertising in many companies.

As long time display channel advertisers know, display is very broad and covers a lot of different types of media inventory, buying sources as well as endless targeting options. So it is certainly a channel that can be utilized for any type of purpose or business goal.

Given this array of targeting options, what are the main strategies that successful display advertisers employ for their marketing campaigns?   Some of the top strategies include:

1. RETARGETING
For the last 2 years, every company from Google to small agencies have been promoting retargeting for their marketing campaigns quite simply because it works almost every time. Essentially, retargeting involves targeting visitors to a client’s website (usually who didn’t take the desired action of filling out a form, buying a product, etc.) and serving them display ads as they browse around the web on various sites.  The basic premise being that because of their prior visit, this audience group has a higher likelihood than the average person to eventually take that desired conversion action.

Figure 1: Retargeting (Site) Example Workflow


2. VERTICAL/DEMOGRAPHIC/SITE TARGETING
This form of display targeting in principal works in a similar manner to how a TV or print advertiser decides which television network and show or news publication to buy media advertising.

If you are offering a finance product, chances are that your messaging appearing on publisher sites like Yahoo! Finance or Bloomberg will have a greater likelihood of resonating with an audience than a lifestyle site like Babycenter. This extends in a similar vein when you are talking about demographics such as age, gender, income levels, etc.

3. 3^rd PARTY DATA
One of the limitations of retargeting is that your volume is limited to the audience that has visited your site. Thus, the only way to grow this volume is by increasing traffic to the site. Many companies have sprung up in recent years in part to help fill this void by selling data to advertisers and their marketing/technology partners to increase the qualified pool of people to target. There are many forms of this data, but a couple of the major forms include prior browsing habits on other sites and prior search engine queries.
 
For example, if you are a travel advertiser selling hotel accommodation in Las Vegas and you can target someone who recently searched for that in a search engine or indeed was browsing on another travel site but did not convert, that is a valuable audience to target.
It is important to state that this data is not personally identifiable and does not contain information that an individual would consider private like a social security number, credit card details, home address, etc.   

4. CREATIVE MESSAGING
Another benefit of display advertising, which differs from SEM, is the ability to target an audience directly. Therefore if you have a multi-stage funnel sales process as most businesses tend to have, marketers can message people individually to both the stage of the funnel that the customer is in or tailor to the specific product or service he or she may be interested in.

For example, if you are a retailer who sells red and blue shoes of various brands, you can dynamically serve that particular product or similar products to a consumer who browsed that particular product page on your site. In a similar vein, in a multi-funnel process if the marketer is in the publishing business that is offering a free 7 day trial, the company can serve those that have not signed up for the trial different messaging to those that have but not committed to a full paid subscription.

5. REAL TIME BIDDING (RTB)
Real Time Bidding probably requires a whole series of posts and as Display (and other forms of digital) advertising continues to evolve, will probably viewed less as a strategy and rather the default best practice of how media is bought and sold. Without getting overly technical, RTB is bidding for each individual impression each time a given publisher site loads. When applied with a robust algorithm and advanced platform that is able to bid for that inventory with the ROI criteria for a given advertiser in mind, it allows for each impression of media to be priced correctly given the browser who may be exposed to the creative message.

With all these Display strategies, there is lot of complexity that can be employed to maximize ROI and to avoid redundancy in marketing outlay. Ultimately, not only should these campaigns be optimized correctly and efficiently as standalone marketing efforts, they should be tracked and optimized in conjunction with other online media channels like SEM and social media to ensure the best possible return for ad spend.

Chris Jacob

The Brand Value of Facebook

In 2010, Facebook topped the list of most visited websites in the U.S., and it is estimated that overall worldwide ad spending on Facebook will likely reach more than $4 billion this year and $5.74 billion by 2012. 

Because of the tremendous potential it holds for our clients, Efficient Frontier has been in the unique position of working up close and personal with Facebook engineers and account reps for the past several months to optimize performance and enhance our clients’ user experience. Thus, we bring you the first in a series of blogs about how to maximize Facebook advertising campaigns, supported by our data so far, and the insights gleaned from working closely with the social networking giant. 

How are Facebook advertising campaigns performing? How do they compare to search? Are they effective? Can we integrate them into other channels? In short, the answers are: the campaigns are performing well thus far; in comparison to search, it depends; yes, Facebook is an effective advertising tool and; yes, campaigns absolutely can and should be integrated.

How are Facebook advertising campaigns performing?

For our clients, Facebook is performing very well. Luckily for Efficient Frontier, our clients tend to be smart, savvy, and open to new strategies, especially the dozen or so forging ahead into the realm of Facebook advertising under Efficient Frontier’s management. Staying open to new strategies is important when advertising with Facebook since you need to be able to move away from the direct marketing results that take place on Google and MSN and into more of a branding mindset. This is not search, it’s Facebook.

Yeah, but how does it compare to search?

It depends. The data varies by industry and client right now and there is no set standard for performance. Facebook is not an advertising platform that has been around for 13 years as search has, so that’s okay. On the cost per click (CPC) side, the performance has varied on either side of the norm. We have seen CPCs on Facebook 3X cheaper than on Google, and with other clients, 1X more expensive. The consistent factor is that the CPCs on Facebook have been relatively competitive with those on Google. On the cost per lead (CPL) side, the performance favors Google because it converts better. But it should convert better because users are looking for you on Google, whereas on Facebook, they are not. Though the CPLs are consistently trending higher on Facebook than on Google, that number varies (between 1 X – 7X) depending on the client, though 3X – 4X tends to be the current average.

So is this effective?

Yes. There is a branding value to Facebook that can’t be overlooked. It’s the Friends of Connections (FOC) effect, or more simply, the friends of users who like your page. The effect is highlighted below:

If you only account for direct leads from Facebook, you are missing half the story. Looking at the above graph (which I made up for demonstration purposes), giving 100% credit to FOC leads brings a CPL of $50 down to $16.67, or giving only 50% credit to FOC leads, brings the CPL down to $25. Though we can target connections of a brand, the FOC metric on Facebook is currently not something you can track (although there are betas around the corner). So then, how do you measure what you can’t see? You don’t. But, you move towards the branding mindset to understand that the effect is there, and how it impacts your bottom line. And you leave a little leeway in your CPL goals for it, and look to a more holistic, cross-channel CPL.

Should we integrate this into other channels?

Of course you should. How? Efficient Frontier can help.

Was that too obvious? In all seriousness, being able to dissect and analyze data from SEM, SEO, display, and Facebook, and measure the value/impact they have on one another and your business is paramount right now. Being able to look at a forecast at various budget points and from there allocate costs across search, display, and Facebook, so that you are getting the best bang for your buck overall, is key. We’re doing it right now at Efficient Frontier, and it’s exciting!

Accurately measuring your brand value is just one of many important issues to consider in regards to Facebook advertising. Stay tuned for several more blogs on Facebook topics in the coming weeks.

Karen Maciolek
Senior Account Manager

January 2011 Display Publisher CPMs: Less time to watch You Tube?

As we continue to analyze and publish display trends for CPMs (cost per thousand impressions) from a seasonal perspective, it is interesting to reflect on what we have learned from both effective advertising and user behavior.

Last week, we published the Month over Month (MoM) CPM trends by exchange and there were some mixed signals between the different exchanges. In terms of the types of publishers and their respective MoM trends, the CPMs are more in line with the Google AdX overall trends with CPMs being down MoM.

We note in many of our posts regarding CPMs, that end of quarter prices are often inflated as particularly brand advertisers often rush to spend unused budget, artificially inflating the biddable inventory CPM levels.  Often advertisers may end up with a declining inventory costs comparing a last month of a quarter with a first month (i.e. Dec 2010 v Jan 2011)

Figure 1 shows the different publisher types compared against each other and Figure 2 shows actual numerical performance per Publisher type and how they trended MoM

Figure 1: Efficient Frontier US Client Publisher Type CPMs - January 2011

Figure 2: Efficient Frontier US Client Publisher Type Summary – December 2010 

January 2011 saw a decrease in all Publisher verticals, except commercial sites (a web presence for a specific business which also has ad space). Not only did CPMs decline MoM across the board but so did the CTR (click through rates) and thus the propensity of browsers to interact with the ad creatives.

Interestingly, both the number of sites and impression share of shopping sites increased proportionately slightly in January, possibly indicators consumers are still chasing bargains long after the official holiday season.

One of the most obvious trends occurred with the video publishers, whose CPMs declined 34% in January with a similar decline in their relative Impression share. Combined with the overall decline in CTRs, can we conclude that browsers have less disposable time in January than in the more holiday friendly month of December?

January is traditionally a “back to work/school” month, and many industries - finance, legal, real estate - start to get busier again after the holiday period. So while a December 23^rd day in the office might afford the luxury of an employee checking personal email and forwarding You Tube videos to friends, a January 12 work day does not.

From an advertiser’s perspective, all of these trends are important to note as it relates to designing strategies to garner more volume and broadcasting their message at different times of the year. For example if you are targeting a user demographic that is more inclined to video sites, take advantage of the lower rates on offer.

If you are a direct response advertiser concerned about falling CTRs of your creative messages in January, look at it in the context of the overall season. The maybe test your current messages on gaming or finance sites specifically where the CTRs are more constant to gauge whether a mass creative overhaul is really necessary.

We will continue to look at these trends to give advertisers insights in how to best analyze and optimize their display marketing dollars.

Chris Jacob

What we learned from Valentine’s Day: How retailers can navigate negative effects of “romantic” holidays on Average Order Value (AOV)

At Efficient Frontier, we love to tease out patterns and use data to inform decisions. One pattern we’ve noticed is that the Average Order Value (AOV), or the average dollar amount spent for each customer order, tends to drop for retailers during the peak Christmas shopping season. One theory to explain this unlikely phenomenon is that consumers who purchase gifts for others tend to buy less expensive items than if they were shopping for themselves. We can also surmise that certain people only become customers during certain seasons, so the mix of customers changes.

The graph below shows the AOV trend for two groups of clients. Group A contains retailers that skew toward romantic and sentimental items commonly gifted during Christmas or Valentine’s Day. Group B contains retailers that tend to sell more practical wares that are not typically considered romantic. We can see from the graph that both groups dip in AOV during the Christmas shopping season. However, Group A also dips to mid-December levels beginning around the third week of January. Group B also takes a slight dip beginning on the 1^st of February, just before Valentine’s Day.

That’s not to say that order volume doesn’t play a role. It most certainly does, though the volume trends are not quite the same.

Besides the obvious fact that not all retailers have the same seasonality trends, we can see some similarities that arise across the board. So why does any of this matter?

Whereas Christmas has a well-defined holiday shopping “season” (beginning on Black Friday), other holidays, such as Valentine’s Day, don’t have such clearly defined shopping seasons. A change in AOV can indicate a change in the mix of customers. To capitalize on this, retailers can tailor ad copy towards customers who are seeking to purchase gifts.

Also, since only Group B showed a spike during Thanksgiving when there were a lot of deals, this suggests that retailers should experiment with NOT discounting romantic goods by much.

In addition to ad copy changes, campaigns need to be set up to capture the increase in demand. This means dynamically changing budgets and bids with the appropriate spend mix on high and low AOV campaigns. All of this could help maximize conversions from customers who are planning to buy gifts during the holiday season.

Based on this data, we also recommend that retailers:

• Use historic data to anticipate demand, and pace budgets accordingly
• Experiment with pricing on less time-sensitive but essential retail goods, which show a high degree of elasticity. “Romantic” goods are time-sensitive, and do not usually capture additional demand from low pricing during the high season
• Align your campaigns in a cross-channel fashion. Use Facebook for upper-funnel optimization to generate demand early on in your holiday campaign, and get aggressive on search to capture lower-funnel intent late in the campaign

Ultimately, it all boils down to predictive modeling and proactive account management. In order to maximize your ROI, you need best-of-breed adaptive learning algorithms to place the right bids at the right times. You also need a world-class team to provide human insights and provide the right creative elements.

Joe Romero

Business Analyst

The Valentine's Effect in Display

A lot of people in society both in relationships and single are very cynical of Valentine’s Day as a commercially manufactured day designed to drive sales of flowers, chocolates, cards, jewelry, etc. While that is an argument for another day, we wanted to see if consumers’ behavior online changes as a result of Valentine’s Day and thus give advertiser insights into marketing through this period.

To this end, we looked at Valentine’s Day and the surrounding period in 2010 to see whether Display Publishers saw any traffic variations and whether advertiser creative on these sites tends to be more effective at capturing people’s attention.

Figure’s 1 and 2 below show 3 Publisher sites; one popular social networking site and two well known dating sites and their daily volume and Ad CTRs (click through rates) over 4 periods;
a) Feb 8^th – 12^th Feb (week before Valentine’s Day)
b) Feb 13^th (day before Valentine’s Day)
c) Feb 14^th (Valentine’s Day)
d) Feb 15^th – Feb 19^th (week after Valentine’s Day)

Figure 1: Publisher Traffic Volume from EF US Advertisers 

Figure 2: US Advertiser CTRs (click through rates) on Publisher Sites

Last year Valentine’s Day fell on a Sunday, meaning the 8^th-12^th period and 15^th-19^th period was a Monday to Friday and of course the day before was a Saturday.

It is evident from there is a definite spike in traffic on the 13th which holds into Valentine’s Day. This then drops sharply the following week at a level of around 20% for the Social and Dating A side and over 12% for Dating B from the Valentine’s weekend level. For Social and Dating A, the traffic volume was a similar level below the previous week’s (Feb 8-12) traffic level.

Now while it is probably not surprising that these types of sites garner elevated traffic levels in the lead up to and on Valentine’s Day itself, it is also interesting that all sites also recorded higher CTRs for ads appearing on their site. Bear in mind the ads on the site were not necessarily anything to do with Valentine’s Day, dating or indeed anything social and could have been advertising a Financial product.

Often in both Display and Search, elevated traffic levels lead to lower CTRs but from this sample set in 2010, we show that Display advertiser’s regardless of their vertical can benefit from ad placements on Social and Dating sites during this time.

The data may also indicate that our loud cynical statements about Valentine’s Day may not match our actions. Love is certainly in the air online!

Chris Jacob

January 2011 CPMs: Mixed Trends

Just as Efficient Frontier regularly does with search CPC (cost per click) trends, we have decided to begin reporting on display CPM (cost per thousand impressions) trends.

Figure 1: Display Trends
 

In January, CPM levels on Display followed search CPC trends and are down Month Over Month (MoM) but more similar Year Over Year (YoY).

In display, it is often the case that CPM levels become more volatile at the end of quarters as many Display advertisers have to adjust their spend to meet pre-determined Quarterly budgets. Additionally a lot more inventory is sold on a guaranteed basis in December, particularly the more premium placements, meaning less of that type of inventory available in the biddable market.  All of this makes CPM levels a lot more varied and combined with the fact December is a low season for traditionally larger Display spend advertisers like Finance, this may be influencing the large MoM differences.

Direct response advertisers have to be more vigilant and sophisticated in optimizing their campaigns during these periods as they experience the unique nature of the Display marketplace. We would expect CPM levels to continue to rise through Q1 2011 with more volume from current display advertisers and newer entrants to biddable display continuing to enter the market.

Chris Jacob

Do we surf while we watch TV? The Superbowl Story

My friends call me a dinosaur. “Everyone”, I am told, surfs the Internet when they watch TV. Not me; I watch TV and surf at different times. If the image of most consumers watching TV with a laptop in front of them is indeed true, the implications are significant for multi-channel advertisers.

To test the hypothesis, I looked at paid search traffic during Superbowl. I looked to see if website traffic trends mirrored events on TV. The verdict? Mixed. While there is some interaction between the two channels, it is not as obvious as this. Traffic dipped slightly during the Superbowl kickoff and then built up as time went on. It dropped again at the start of half time but increased at the end of it. Traffic in the second half remained significantly higher than in the first half. Did most people, realize by then that the game was already won? Traffic again dropped during a play call challenge and then finally when the game was over.

This analysis gives us a few insights into how people surf.

(1)    On big TV days, people do surf the Internet while watching TV. While the connection is not very tightly coupled, it will vary significantly by geography. The site traffic from Green Bay on Sunday would have strongly reflected the progress of the game. The same, one surmises would not be the case in Burlington, VT.

(2)    Online and offline activity are often coupled in counter intuitively. I expected online traffic to increase when the game winded down. The opposite happened. Further, people surfed less when half time started and increased as half time went on. Were the Black Eyed Peas so bad?

Multi channel advertisers should take note of these trends. A TV advertisement is no longer an island. As the data shows, it can affect online activity. If the goal is brand building and mind share then ads must invite users to engage with their sites online. When doing media buys on Superbowl or other prime TV spots, advertisers will be well advised to look at online traffic for indications of effectiveness.  Marketers ideally wants spots where there are short dips in online activity followed by big upward spikes as this indicates a temporary attention grabbing event on TV and then an increase in surfing behavior.

What are your thoughts ? Do let us know.

Dinosaur. Siddharth Shah
Sr. Director, Business Analytics

 

January CPCs: Seasonally down but up YoY

CPCs in all major sectors dropped on a Month over Month (MoM) basis after the strong holiday season. As expected, Retail CPCs dropped the most MoM by over 27%.

However,  CPCs were up in all sectors YoY. With a 21.3% ncrease, the travel sector saw the biggest rise followed by the automotive,retail and finance sectors that saw increases in the 6.5-7.7% range. Unlike 2010 we do not expect to see big YoY CPC increases in 2011.  Yet, the YoY CPC trends for January indicate an increasingly competitive marketplace with advertisers continuing to increase their budgets with the economic recovery.

Dr. Siddharth Shah
Director, Business Analytics

The value of a Super Bowl ad: A performance marketer’s perspective

Super Bowl ads are a brand marketer’s dream. Last year’s Super Bowl set a new record for the most-watched program in television history, bringing in a whopping 106.5 million viewers across the U.S. Major brands including Google, Coke, GoDaddy, and several automotive and retail companies bought millions of dollars in advertising spots during the game to capitalize on this massive market.  

Yet the value of TV ads is a subject of much debate among performance marketers. While many marketers believe that TV ads improve marketing performance in the long run, others believe that they are an expensive way to build awareness. At Efficient Frontier, we have attempted to answer this question for many advertisers, and since the Super Bowl is only four days away, we thought it would be an opportune time to present some of our research on the connection between TV advertising and Search Engine Marketing (SEM).

First, does TV advertising help build online traffic? The answer is, undeniably, yes.

The chart above shows the brand impression volume for eight weeks during a TV flight. One can clearly see that the impression trend is very closely tied to the trend in TV spend. A closer inspection reveals that the connection between the two is more nuanced. A well established brand with a high degree of awareness will see a smaller jump in online traffic during a TV flight than a small brand. Furthermore, a big TV ad campaign will lead to a bigger impression spike than a small one. As a rule of thumb, a TV flight causes between a 60 and 80% jump in searches for the brand terms and between 40 and 60% for generic terms.

Super Bowl ads provide a huge spike in awareness. Does the spike translate to sustained online traffic? Some of our research attempts to answer this question. Using vector-auto-regressive models, we were able to estimate that a big spike in TV spend increases SEM traffic for four to six weeks.

 

For all Super Bowl advertisers, we advise the following to maximize your ads’ effectiveness online:

1. Anticipate a spike in traffic volume on your site both from paid and organic searches. Ensure that both your brand and non-branded terms are at high positions to capture all the traffic.

2. While you may be worried about spending beyond your allocated SEM budget on Sunday and perhaps for the coming days, you will also see much better than average conversion rates. This is good as it means lower acquisition costs.

3. Ensure that your Facebook campaigns are aligned with your TV efforts and link your Facebook pages to the video of your Super Bowl ads. Measure the effectiveness of your Facebook campaign using upper funnel metrics such as average time spent and engagement, and less to lower funnel metrics such as conversions.

4. Measure the effectiveness of your Super Bowl ad on your online marketing efforts right after the ad runs, then again after a week, a month and beyond. The results may pleasantly surprise you!

While our analysis of TV and SEM data has provided us with valuable insights, it also opens the door to tantalizing media mix questions. What is the right mix of online and offline budget to maximize a marketing campaign’s performance? How should the online budget be spent when TV flights are known in advance? What are the long-term branding effects of a TV campaign and how can they be measured online? What is the effect of TV advertising on Facebook, which is an upper funnel marketing vehicle? We do not have all the answers yet, but we are working on finding them and we’re very excited about the possibilities.

Dr. Siddharth Shah
Director, Business Analytics