Connect with FOLIO:
         

ADVERTISEMENT



FOLIO: Personalities -- The Blog People Page


Matt Shanahan

Prospecting Up-sells with Behavioral Analytics

Matt Shanahan emedia and Technology - 01/04/2012-12:34 PM

 

This post is republished with permission and originally appears here.  

In paid content, one of the challenges for corporate sales is finding demand for content that can be monetized – finding a good lead. Our research shows one of the best sources for leads is within existing individual subscribers where several individuals are sharing the access to the paid content. The charts below show a typical example of how to identify an individual subscriber that is a lead for an up-sell to a corporate or group agreement.

The first chart shows the daily use profile for two different individual subscribers by graphing the total number of reports accessed over a 90-day period. Each daily use profile is color coded based on the unique devices used to access the reports. The daily use profile with just the blue color shows how one subscriber’s account accessed reports via a single device. The daily use profile with multiple colors shows how the other subscriber’s account accessed reports via five devices (each color representing a distinct device).



Whereas the daily use profile with one color represents a loyal subscriber, the daily use profile with five colors is a prospect for an up-sell because of unmonetized demand of multiple users sharing one subscriber’s account. But how can you be sure the multi-color profile isn’t simply a raving fan?

The number two is a good start. In the hourly use profile for each account, the shared account has twice as many active hours as the individual account. Additionally, some devices are active earlier than other devices (e.g., purple vs. orange) but overlap on their activity which means they are used in different time zones. Also, note the individual account profile shows inactivity for lunch in the middle of the day. There are no low activity hours for the shared account rather it peaks at the periods of most overlap between devices.



Looking at the quarterly use profile, it becomes clear that each of the shared account devices were active throughout the period. Unlike the individual account which had ten inactive work-days during the period, the shared account had no inactive days.  Digging deeper into the quarterly profile, the shared account consumes three times the content as compared to the individual account.



The Implication

The shared account exhibits usage profiles with monetizable demand for content. Corporate sales has a number of avenues to pursue the lead either directly, through procurement, or even through compliance. Properly engaged, this lead will convert with high probability.

 

More...
Matt Shanahan

Click-through Rates: The Metric for Missed Expectations

Matt Shanahan emedia and Technology - 09/13/2011-12:23 PM

This post is republished with permission and originally appears here. 

Click-through rate (CTR) is often used to describe the advertising performance on a publisher's site. CTR for an ad is defined as the number of clicks on an ad divided by the number of times the ad is shown (impressions), expressed as a percentage. If the ad sales team for a publisher claims 1 million monthly unique visitors with 4 million page views and a CTR of 0.2% (or 8,000 click-throughs), the buyer might think those click-throughs are all distributed across the million unique users to yield 8,000 unique conversions. The buyer and the seller are wrong.

Here's the problem: CTR doesn't take into account audience engagement, not to mention the fact that other advertisers are competing for the click-through on the same page. To demonstrate the weakness of CTR as a performance metric, let's assume a single advertiser buys the entire inventory for the month. The advertiser buys all 4 million page views from the 1 million uniques to have an expected 8,000 unique visitors click through to the advertiser's site. But two factors skew actual unique conversion from the CTR's expected unique conversions: CTR differences between visitors and click-through qualification.

• Scout Analytics research shows that the advertiser's desired conversion target, the loyal audience, usually comprises 20-30% of the visitors who generate 70-80% of the page views. This means that the available audience for the advertiser is really 200,000-300,000 visitors not 1 million. The remaining 700,000 audience members - those generated from search - are irrelevant to the advertiser.

• Scout Analytics research also finds that click-through behavior varies by engagement, with loyal audience members delivering the majority of click-throughs because of relevance. For a myriad of reasons, the 700,000 irrelevant fly-bys still generate click-throughs, but at a lower CTR. For this example, let's assume their CTR is half that of a loyal audience.

• Finally, Scout Analytics research shows that some audience members are more "clicky" than others - meaning they click on ads more frequently. The clicky audience not only generates click-throughs more frequently, but often do so on the same ad, distorting the CTR further. (This example assumes 10% of the loyal audience is twice as clicky as the average loyal audience member.) Click-throughs from clicky audience members need to be de-duplicated.

After solving a couple of simultaneous equations, the CTR of the average loyal audience is 0.217%; the clicky audience's CTR is 0.434%; and the fly-by CTR is 0.109% The graphic below shows what these CTRs do in delivering conversion rates. The horizontal calculation follows traditional CTR thinking to yield the expected 8,000 unique conversions. The vertical calculation, called Audience Conversion Rate (ACR), follows a methodology of tracking unique audience conversions by qualifying the click-throughs as in-market and unique. Rather than getting 8,000 target audience members to click through, the advertiser would actually get 6,087, a full 24% short of the expected 8,000.

The more relevant ACR metric measures the number of unique audience members that clicked through divided by the total number of loyal audience members. Looking at the above example, the publisher is actually producing 2% ACR (i.e., 6,087/300,000). Not only is the 2% rate more representative of publisher and media performance, it is also a better value proposition to the advertiser. This makes your ad sales more productive.

More...
Matt Shanahan

3,044 Reasons Why CTR is the Wrong Metric for Media Buying and Selling

Matt Shanahan emedia and Technology - 09/01/2011-13:06 PM

Click-through rate or CTR is often used to describe the advertising performance on a publisher’s site. CTR for an ad is defined as the number of clicks on an ad divided by the number of times the ad is shown (impressions), expressed as a percentage. If the ad sales team for a publisher claims a million monthly unique visitors with 4 million page views and a CTR of 0.2% (or 8,000 click-throughs), the buyer might think those click-throughs are all distributed across the million unique users to yield 8,000 target audience members. The buyer and the seller are wrong.

Here’s the problem: CTR doesn’t take into account audience composition. To demonstrate the weakness of CTR to communicate performance, let’s assume a single advertiser buys the entire inventory for the month. The advertiser bought all four million page views from the one million unique to have an intended 8,000 target audience members click-through to the advertiser’s site. What skews CTR to be a poor metric of performance? Engagement.

· As shown by Scout Analytics research, the advertiser’s target, the loyal audience, is usually 20-30% of the visitors who generate 70-80% of the page views. This means that the available audience for the advertiser is really 200-300,000 visitors not 1,000,000. The remaining 700,000 audience members generated from search are statistically irrelevant to the advertiser.

· Scout Analytics research also finds that CTR varies by engagement with loyal audience members delivering the majority of the click-throughs, but for odd reasons, irrelevant fly-bys still generate click-throughs. These 700,000 irrelevant fly-bys typically have a much lower CTR. For this example, let’s assume their CTR at half the CTR of a loyal audience.

· Finally, Scout Analytics research shows that some audience members are more “clicky” than others. The clicky audience not only frequently generates click-throughs but often on the same ad distorting the CTR further. This example assumes 10% of the loyal audience is 5X more clicky than the average loyal audience member.

After solving a couple of simultaneous equations, the CTR of the average loyal audience is 0.177%; the clicky audience’s CTR is 0.885%; and the fly-by CTR is 0.089%. Rather than getting 8,000 target audience members to click through, the advertiser would actually only get 4,956, or 3,044 short of the expected 8,000 – a full 38% short.

A more relevant metric would be audience conversion rate (ACR) or the number of unique audience members that clicked through divided by the total number of loyal audience members. Looking at the above example, the publisher is actually producing 1.65% ACR (i.e., 4,956/300,000). Not only is the 1.65% a more representative of publisher and media performance, it is also has a better value proposition to the advertiser and makes ad sales more productive.

More...
Matt Shanahan

The Only Thing Worse Than a Fly-By…

Matt Shanahan emedia and Technology - 07/05/2011-09:37 AM


This post is republished with permission and originally appears here.

...is a scraper! Or at least scrapers that aren't monetized properly. Scrapers are users or automated services that systematically consume media content - especially news. Their motivations are usually commercial in nature such as a media monitoring service, advertising verification, or lead sourcing. A scraper can be identified by the fact that their volume of consumption is several standard deviations or more higher compared to the rest of the audience - even compared to fans.

It's not uncommon for scrapers to make up less than 0.1 percent of an audience and generate 10-15 percent or more of the page views (i.e., ad inventory and revenue capacity). The graph to the right illustrates the behavioral difference between the largest segment of visitors, fly-bys, and the smallest segment, scrapers. On average, scraper behavior generates 100-150 times more page views than their percentage of the total audience represents.

So why is a scraper bad for publisher revenues? Aren't page views good?

The first reason is that scrapers devalue a publisher's ad inventory by lowering conversion rates for advertisers. Page views delivered to scrapers are like page views delivered to bots. An advertiser that purchases a scraper page view is wasting their money.

Therefore, mixing scraper page views with regular audience page views produces lower conversion rates. As advertisers see the lower rates, they move their advertising dollars to higher performing sites.

The second reason is that scrapers decrease sell-through rates and CPMs. Many publishers do not have 100 percent sell-through rates and as a result, it is harder to negotiate ad rates with advertisers because the publisher is trying to get all the inventory sold. In the case of selling inventory without scraper page views, the publisher creates scarcity along with higher quality (see first reason) to raise prices. Because scrapers reduce sell-through rates and CPMs, they hurt a publisher's revenue.

The third reason is that unmonetized scrapers represent lost revenues. Remembering that scrapers have economic motivations to consume media, publisher can provide a different value proposition. Rather than monetizing their behavior like the rest of the audience, scrapers can be charged for direct licensing of the content. The beauty of direct licensing is that the publisher typically gets a multiple over the ad revenue these scrapers would have produced, and for the ones that don't license the content, they are at least blocked from devaluing the ad inventory and lowering CPMs.

While some publishers try to identify and stop excessive consumption, the bar is often set too high.

More...
Matt Shanahan

Building a Loyal Audience? That's a Business Model

Matt Shanahan Audience Development - 06/16/2011-10:48 AM

This post is run with permission and originally appears here. 

As I pointed out in my previous post, advertisers don't buy page views they buy audience. A publisher's business model has to produce and monetize an audience. So what kind of audience is profitable? A loyal one. Here is the proof...

The Revenue Model

Because audience engagement is the unit of monetization and because each member engages differently, the revenue contribution and profitability of each audience member varies. For example, assume a fly-by audience member generates on average three page views. With a $30 RPM, each fly-by is worth $0.09 in revenue. Now compare that to a loyal fan generating 100 page views a month or $3 of revenue per month (i.e., $36 per year). 

The revenue model can be plotted as shown in the figure above. Charting audience versus their revenue contribution illustrates the revenue model for a publisher. On the left, highly engaged fans contribute good revenue and on the right, each fly-by generates a small incremental amount of revenue.

The Cost Model 

Aside from advertising sales, the two primary costs in digital media are associated with producing audience members namely audience development and editorial. Since a target audience is finite, acquiring new audience members becomes increasingly expensive as the size of the audience grows, due to the decreasing number of potential audience members remaining. Additionally, as the audience grows so does diversity and the need for a broader range of content at a larger editorial cost. The increasing cost to acquire and engage the audience is represented along with the revenue model in the following chart.

The Profit Model

Of course profits are made when the cost model is below the revenue model. For the fly-by to be profitable, the cost to produce the content and acquire the page views of the fly-by needs to be below $0.09. Acquiring an audience of 30M fly-bys per month would only generate $2.7M in revenue per month or a little more than $32M per year but the costs of that are likely to be much higher (e.g., Demand Media cost model).

By contrast, the cost to produce the content and acquire the page views of the loyal fan needs to be less than $3 per month to be profitable. Research by Scout Analytics shows the importance of a loyal audience to profitability in a revenue model. A loyal audience is made up of all visitors minus the fly-bys and usually constitutes about 20 percent of the unique visitors but are responsible for about 80 percent of all page views. This means that 80 percent of the revenue capacity (i.e., ad inventory) comes from a loyal audience while about 20 percent come from fly-by visitors. So even though fly-bys make up the vast majority of unique visitors, their revenue contribution is astonishingly low. In other words, generating revenue from fly-bys is the same as chasing page views, but building loyal audience is building a business model.

 

More...
Matt Shanahan

Counting Page Views? Don't Call It A Business Model

Matt Shanahan emedia and Technology - 05/26/2011-08:30 AM

This post is reprinted with permission and originally appears here.

Measuring the number of page views as a key performance indicator (KPI), is a growing practice among publishers. In fact, editorial and development teams are increasingly being rewarded for boosting page views, with some publishers even shaping their entire site just to generate page views.

That is not a business model! Let me prove it with an extreme example.

Any publisher can deploy bots to generate page views for their site. No advertiser will pay for those page views, because the page views have no advertising value. While page views could be used as a KPI by the editorial team to generate more content for bot consumption, no revenue is coming through the door to keep them employed.

The right metric for publishers should be revenue performance indicators (RPI), which means the metrics tie directly to the business model. Many publishers are looking to build recurring revenue streams from loyal audience members, and in this case, RPIs such as audience size, loyalty, and level of engagement are meaningful. However, some publishers are relying on non-recurring revenue from SEO acquired visitors, and in this case, RPIs such as percentage share of search and time on site become more relevant. In paid content, RPIs such as price per article or price per device become critical. And for all of these business models, average revenue per user (ARPU) is the RPI for benchmarking efficiency and profit (see my post on ARPU here).

While correlating user behavior to the business model is the only way to judge revenue performance, surprisingly few publishers can differentiate between a page view that is aligned to the business model vs. a page view that is not. Consequently, many publishers are chasing low value page views and jeopardizing their long-term viability. Might as well hire some bots.....

More...
Matt Shanahan

What About an Ad Wall?

Matt Shanahan emedia and Technology - 02/15/2011-14:26 PM

This post is reprinted with permission and originally appears here.

In the pursuit of earnings, publishers have done two things - intensified SEO efforts, which creates fly-by ad revenue and established paywalls, which create fan subscription revenue.

Remember these definitions:

• Fly-by - Someone who visits once.
• Occasional - Visitor to the site two to three times per month.
• Regular - Visitor to the site one to two times per week.
• Fan - Visitor to the site more than two times per week.

Some fans are willing to subscribe through a paywall, and fly-bys are commodity impression to be bought and sold through exchanges, but what about the other 15-25 percent of the audience?

The barbell focus on fans and fly-bys for revenue optimization overlooks a significant untapped opportunity in between the two groups. The regulars and occasionals as we call them represent an opportunity for publishers to create differentiated advertising products for advertisers such as an adwall.

Users have to give time to get time, that is the underpinning of attention economics and online advertising. Advertisers underwrite user access to media or other digital goods in exchange for a moment of their attention.

I was reminded of this exchange on a recent flight from New York. After take-off, I decided to use the in-flight Internet access which normally costs $9.95 for that route but was being offered for free if I would watch an advertisement. For me, the exchange of 30 seconds of my time for four hours of Internet access was a no-brainer. In gaming, companies like WildTangent are allowing users to watch advertisements in exchange for virtual goods to extend their playing time online. In exchange for watching the latest episode of a TV show, the Hulu user watches advertisements.

These are all examples of adwalls where there is an exchange for watching an advertisement, i.e. the user is given access to digital media, games, etc. Unlike interstitials that are frequently dismissed, the adwall requires viewing and acknowledgement or no access is provided. Adwalls are intended to optimize revenue from the regulars and occasionals that make 15-25 percent of the audience. To illustrate the idea, let's consider a couple of different scenarios of how fly-bys might react to an adwall vs. regulars.

The first scenario is Demand Media. Demand Media is a company built on SEO. The Demand Media audience is mainly fly-bys. How much is an article on "how to window shop" worth? Would a fly-by be willing to watch a 30 second advertisement in exchange for access to the article they are likely to look at for at best a couple of minutes? For most people, the answer is likely to be no. In this case, the adwall turned away traffic that could have been served by commodity ads from an exchange.

The second scenario is The Times of London. While some loyal fans have signed up through the paywall, many others have not. Surely, many non-subscribers and previous readers would be willing to watch a 30 second advertisement for a one day pass. In this case, the adwall extended the revenue stream that otherwise would have been a lost opportunity.

Finally, ad units like adwalls are unlikely to erode any existing paywall revenue because the fans that would pay for access would also likely pay to avoid the adwall for convenience sake. It actually has the chance to generate paywall revenue. As our CTO stated, he would gladly pay Google $10 a year to get rid of the ads in YouTube.

This post is less about the merits of an adwall per se and more about the merits of targeting higher value ad units to engaged audience members to optimize revenue. An adwall without targeting would most likely end like slate.com. The slate.com adwall probably failed because it did not target the adwall to certain audience members or around specific differentiated content. Consequently, the give-to-get equation was out of whack.

Digital media cannot rely on a paywall to generate all the earnings for shareholders. Advertising needs to generate more revenue and an adwall is an example of just that. It creates a differentiated product for advertisers, a reasonable value of exchange for the audience and incremental revenue for the publishers.

Matthew Shanahan is the SVP of strategy for Scout Analytics, which creates actionable revenue opportunities for digital publishers by tracking and
targeting user engagement.

More...



CONNECT WITH FOLIO: NOW
         


CAREER CENTER dots icon