Will Scully-Power

Facebook Sponsored Stories Ads Have 46% Higher CTR, 18% Lower Cost Per Fan Says TBG Digital Test
May 4, 2011, 11:25 pm
Filed under: #mktgcloud, Datarati, Social Media | Tags: ,

Facebook Ads API service provider TBG Digital has revealed that in a 10-day, 3-client, 2 billion impression test, Facebook’s new Sponsored Stories ad units received a 46% higher click through rate, a 20% lower cost per click, and an 18% lower cost per fan than Facebook’s standard ad units.

The new ad unit has looked especially promising because it so tightly fits in with how people use Facebook. It converts a user’s activity, such as Liking a Page or a using an application, into a sidebar ad shown to their friends. Those friends can then respond to the ad unit, through taking actions such as Liking the Page or playing a game.

With evidence like what TBG is providing, advertisers should strongly consider adding Sponsored Stories to their ad mix, and Ads API tools and services should move quickly to include the ad type in their offering.

TBG Digital carries out the purchase of up to 2 billion Facebook ad impressions a day via its Ads API service for big name clients including JetBlue, American Express, and Vodafone, making it one of the largest ad buyers on the platform. When Sponsored Stories launched in January, it moved quickly to be able to provide purchases of the ad unit at scale to its clients, and has integrated thethree new Sponsored Stories types released last week into its offering. While Facebook reported that its own tests showed increased brand lift, engagement, and ad recall, there wasn’t conclusive evidence of how much better Sponsored Stories performed than standard ad units until now.

The three clients involved in the test were a retailer, a consumer packaged good brand, and a game developer, ensuring a breadth Sponsored Stories types and destinations were included. TBG Digital’s CEO Simon Mansell told us the company took care to keep the Sponsored Stories and standard ad groups as similar in targeting as possible preserve the integrity of the test.  These facts combine with the high impression volume of the test to make the conclusions reliable.

Ads API-based media buyers such as TBG Digital test thousands of ad creative and targeting variants per campaign to attain the highest click through rate possible. The creative and targeting of Sponsored Stories are generated automatically, meaning Sponsored Stories not only represent a massive increase in click through rate, but a reduction in labor on the media buyer’s side.

Now that many brands have established their Facebook presence, they are concentrating on gaining fans. It now appears that the several Sponsored Stories ad types designed to drive traffic to Pages (Page Like stories, Page Post stories, Page Post Like stories) can reduce the acquisition cost per fan by 18%. This is a significant cost reduction considering many brands are looking to secure millions of additional fans.

Since Sponsored Stories don’t depend on media buyers discovering lucrative niche demographics with high click through rates that Facebook charges more for, Sponsored Stories campaigns have a 20% lower CPC according to the test. This could translate into big savings for brands, performance advertisers, and especially game developers who look to bring in millions of new players with ad campaigns supporting the launches of new games.

The data from this test support the assumptions some in the advertising industry have made about the shift towards social ad design. Sponsored Stories ad units take advantage of Facebook’s inherent strength — the social graph — to offer ad design unavailable in traditional search and display. Mansell tells us that currently about 10% of the impressions it buys are through Sponsored Stories, but that he expects that to increase to 25% as more of TBG Digital’s clients are signing on to use Sponsored Stories every day.

Since standard ad units are necessary to seed the initial actions that trigger Sponsored Stories, advertisers will always need to buy ads that don’t depend on user behavior. But data indicating that Sponsored Stories can deliver more clicks and more fans for cheaper, expect usage of Facebook’s social ad units to drastically increase this year, and more ad spend to shift from display to the social network.

More: http://www.insidefacebook.com/2011/05/03/sponsored-stories-ctr-cost-per-fa


30 Reasons to Use AB Split Testing for Conversion Optimisation
May 4, 2011, 4:33 am
Filed under: A/B Testing, Datarati | Tags:

From the team at WiderFunnel.

In case you need some more reasons to get started right now, here are 30 reasons you should start using A/B Testing for Conversion Optimization:

  1. Conversion Rate Optimization does not exist without controlled, statistically valid testing
  2. A/B testing finds out how people actually act rather than just the way they think they will, which is all you get from focus groups, user testing and usability testing
  3. A/B testing confirms or disproves the hypotheses gathered from eyetracking, click heatmaps, surveys and all other qualitative data gathering
  4. A/B testing is quantitative and the probability of error is known (and when was the last time your other marketing tactics told you that?!)
  5. A/B testing allows for testing Radical Redesign variations, which can be a great starting-point for most landing pages and site templates
  6. You can test different layouts of site-wide page templates on dynamic, CMS-driven pages with A/B testing (see example)
  7. A/B testing usually delivers more dramatic conversion rate improvements than possible with Multivariate testing (MVT)
  8. A/B/n testing is not limited to just two variations
  9. A/B testing allows you to isolate individual hypotheses to build your “best practices” Playbook
  10. You can’t imagine how bad your web pages and site templates really are until you start A/B testing them
  11. A/B testing allows entirely different messaging approaches that can lead to strategic marketing insights
  12. Insights from A/B testing can be used to improve your other online and offline marketing messages
  13. You can track multi-channel conversions with A/B testing including online lead generationaffiliate revenuee-commerce, and phone call conversion optimization
  14. A/B testing allows for easier in-depth analytics than MVT so you can develop further test hypotheses for follow-up test rounds
  15. A/B testing allows alternative conversion funnel tests
  16. A/B testing allows condensing and expanding multi-step forms
  17. A/B testing can be used for landing page optimization, website conversion optimization, email testing, and Evolutionary Site Redesign (ESR)
  18. A/B testing can be used to run a customized MVT test using any test design
  19. A/B testing allows you to isolate only interesting potential interaction effects without wasting traffic on the inconsequential
  20. A/B testing will not limit your test design options like MVT does
  21. A/B testing gets results faster than any other method while MVT is greedy for traffic
  22. A/B testing does not tempt you to end the test before statistical confidence has been achieved as much as MVT
  23. You can easily prune under-performing variations in A/B testing without re-starting the Test Round
  24. A/B testing is easy to spell
  25. A/B tests can quickly be re-run and confirmed for extra assurance
  26. A/B testing results make both the CMO and the CFO happy
  27. The business results from A/B testing look great on your resumé
  28. A/B testing allows you to test your Marketing Strategy and core Value Proposition
  29. WiderFunnel guarantees your results from A/B testing (and Multivariate too, of course)
  30. A/B testing is fun!
More: http://www.widerfunnel.com/conversion-rate-optimization/30-reasons-to-use-ab-split-testing-for-conversion-optimization

Radian 6 Insights Overview
May 4, 2011, 4:29 am
Filed under: Datarati, Social Media, Video | Tags: , ,

Adam Mosseri from Facebook presents: Data Informed, Not Data Driven
May 4, 2011, 3:58 am
Filed under: A/B Testing, Datarati, Social Media | Tags:

%d bloggers like this: