When you hear the word conference, what comes to mind?

If you’re an online marketing professional, likely you’ll think of Search Engine Strategies, Webmaster World PubCon or ad:tech – and with good reason. Industry events such as these are heavily promoted, well funded and oft-cited in the blogosphere.

However, there’s another conference type that isn’t heavily promoted, isn’t usually well-funded and yet can provide one of the richest venues for seeing the cutting edge of research, participating in the generation of new ideas and building relationships between the too often disparate worlds of theory and application.

I’m talking about academic conferences.

Wait - aren’t those just for grad students and researchers?

Nope – or at least not the one I recently attended.

Chaired by BJ Fogg, head of Stanford’s Persuasive Technology Lab, the Persuasive Technology 2007 conference in Palo Alto, California brought together the world’s leading experts on creating computing products and interfaces to change people’s beliefs and behaviors.

Designed to give attendees "a stronger grasp of persuasion psychology (the big ideas) and new insights for designing video games, mobile phone apps, and web sites to motivate and influence people (the practical stuff)," this conference not only provided a venue for idea presentation, but also gave attendees the opportunity to participate in idea generation.

Through a unique conference format, each of the 40+ presenters had only 10 minutes to talk about their recent work and research. Naturally, 10 minutes to talk about subjects like using digital images to persuade viewers or the way peers influence others online doesn’t leave a lot of room for elaboration.

Fortunately, each speakers’ panel was followed by a question and answer session as well as a break in which all attendees could mingle together, ask questions, generate ideas and begin building relationships.

Unlike some industry conferences I’ve attended, Persuasive 2007 was deliberately kept small (less than 150 attendees) to facilitate a high level of interpersonal interaction. And what’s more, there was nearly a 50/50 split between industry and academic representatives, with a large percentage of attendees from international organizations.

And wow, what research.

Take Jeremy Bailenson’s work at the Virtual Human Interaction Lab on the implications of avatar representation for social influence. One of his team’s projects determined that people are more likely to vote for a presidential candidate that looks more like them than one who does not, without ever being aware of the similarity. Such research has significant implications for the alteration of facial images in the media, both online and offline.

(Read the full article…)

The ability to perform A/B tests on ad descriptions and landing pages in paid search is truly a beautiful (and powerful) thing. If you are not testing ad descriptions against one another for conversion and ROI, then you are missing out on a huge opportunity to gain more sales and leads for your company within your given budget.

However, a lot of times it is tempting to throw 4, 5, 6 or more different ad descriptions into the search engines’ systems and test them all at once. Yet the key to drilling down to conclusive evidence of what does and does not work in terms of driving more leads or sales to your website is testing 1 element at a time. There IS a reason why they call this A/B testing – not A/B/C/D/E/F testing!

Take this example. Say you set up 3 different ads for testing in Google AdWords:

Suppose that at the end of 2-3 weeks the 1^st ad significantly outperforms the others in terms of conversion rate. Was it the ‘5 Diamond’ designation? Was it the fact that the rate was included? Was it that it stated ‘downtown’? You really can’t know for sure because no single variable was isolated. Each ad differs too much from the others and therefore you simply can’t draw any great conclusions from this test.

Instead, what if you were set this up on a ‘testing schedule’ of sorts. For testing in Week 1-2 (assuming this is enough time to gain actionable data… but definitely run the ads for longer if needed):

(note that the only variable tested is the use of a room rate in ad description)

At the end of Week 2, say the first ad with the rate emerges as the clear ‘winner’. You can now conclude that the inclusion of rates in the ad description is a factor that contributes to more online sales.

Week 3-4? Stick with the rate in ALL descriptions now that you know it is important for sales, but move on to a new isolated variable. Say that this particular hotel has a number of honors to work with as key differentiators and now you’d like to see which resonates best with searchers:

Week 5-6? Perhaps the hotel wants to test the use of special offers such as ‘Free Breakfast’ v. ‘4^th Night Free’. Week 7-8: maybe 2 different landing pages.

By patiently remaining systematic in this approach and letting the numbers provide the proof, you’ll have solid evidence of which ad elements and landing pages produce the highest ROI for your company.

One important last word of note though! In Google AdWords, much to the conversion marketer’s chagrin, Google will by default start delivering the ad that has the highest click-through rate – not the ad with the highest conversion rate. If your goal for paid search is to drive the most leads or sales, not the most clicks, then I HIGHLY recommend that you turn off the setting that so-called ‘Optimizes’ your ad distribution.

To do so go to your Google Account and:

1. Select ‘Edit Settings’
2. Under Advanced Options select the radio button to ‘Rotate: Show ads more evenly’

Happy Testing!