Eric Enge Interviews Tim Ash
Published: March 10, 2008
Tim Ash is the President of SiteTuners.com, a leading performance-based landing page optimization firm. During his 14 year involvement with the Internet, Tim has worked with American Express, Sony Music, American Honda, COMP USA, Harcourt Brace, Universal Studios, HomeGain, Fair Isaac, TransUnion, Rand McNally, Red Envelope, Black & Decker, and Coach to develop successful Internet initiatives. He is a highly-regarded speaker many international conferences including Search Engine Strategies, Affiliate Summit, and Internet World.
Tim is a contributing columnist to several publications including SearchEngineWatch, Website Magazine, DM News, and Electronic Retailer Online Strategies magazine. He received his B.S. from UC San Diego "with highest distinction". He also completed his M.S. during his PhD. studies in computer science at UCSD, specializing in Neural Networks and Artificial Intelligence. He is the author of the bestselling Amazon.com Computer & Internet book Landing Page Optimization: The Definitive Guide to Testing and Tuning for Conversions (John Wiley & Sons Press, 2008).
Eric Enge: Let's start with a brief definition of Landing Page Optimization.
Tim Ash: Landing Page Optimization is the science of testing different versions of your landing page to improve their efficiency as contrasted with just designing a website, or redesigning it. It's different from usability, and it's different from web design. When you have an existing page, it's the science of testing different variations of that page to get the best results.
Eric Enge: How is it different than usability?
Tim Ash: In both web design and usability you have somebody that makes essentially one pass through it. They implement their best thinking about good usability, or a good web design, or good graphical design, or good copywriting. That's basically one person's opinion, or the collective opinion of a company, or agency, or your team.
But, it doesn't really consider the input of the most important constituency of all, which is the intended website audience, or landing page audience. So, testing is basically looking at what your audience cares about, not what you think is best.
Eric Enge: Right. So, you could conceive of doing usability work as something that takes place before you launch the website and potentially some follow-up work afterwards, but it's still an attempt to get your first best guess.
Tim Ash: Exactly. That's your baseline that you are going to compare future performance to, and the thing that you test. The things that you propose to put into your test plan can involve changes to usability, and site organization, graphic design, copy, all of that. It can also include the call to action maybe within the offer; all of those things are fair game for what you test, but it's the fact that you are testing many, many different alternatives that matters. You don't really know what's best until you test it.
Eric Enge: Right. Why should you do it? To start that discussion, the way I write about this when I write about Landing Page Optimization is, because the chances that you made the perfect guess the first time are basically zero.
Tim Ash: Yes, that's certainly true. The consequence of that is that you are leaving huge amounts of money on the table. It really goes back to why we are in online marketing. We are trying to be successful, and the most neglected part of that process is the landing page itself.
Typically, you have a lot of thought that goes into driving traffic to your site, optimizing your pay per click campaigns where you are trying to manage them for return on investment or ROI on a keyword level. You have sophisticated email campaigns that continue to give people progressively deeper exposure to your company and hopefully enhance the lifetime value of their relationship.
But, in the middle you have this split second, when they get to your landing page and they either do something you want them to, or they don't. And, that part has been neglected. This is not a slam on graphic design, but, one person, often a junior person, that's responsible for coming up with this page, and yet the whole economics of the business depends on the efficiency of that page.
Eric Enge: You don't want to just throw that money away, is that what you are trying to say?
Tim Ash: Yeah. I don't know where else you could get maybe 10%, 20%, 50%, or even a 100% improvements in your revenue that are long term if not permanent just by looking at your own site. You can't get those kinds of efficiencies out of the pay per click campaigns, because they are an auction model as you know, and it depends on what your competitors are paying. So, you have to pay market rates there; but here you can tweak a few knobs, test the few different pages and come up with a huge advantage.
Eric Enge: Indeed. So, when you talk about pay per click campaigns for example, obviously if you can increase your conversion rate by 30% say, suddenly your ability to up your spend and drive more volume is significantly improved.
Tim Ash: Yes. And, as you know the lion's share of exposure and traffic goes to the very top positions in search results. So, if you can crack into the top three positions or four positions as a result of improved economics, you will get often an order of magnitude more traffic, not just proportionately more.
Eric Enge: Right. So that's very cool. Let's talk about the kinds of things you do during Landing page optimization. What you start with is a webpage or multiple web pages. You decide to change certain things, and you measure how the new version or new versions compared to the original version by running them essentially in parallel sequences; and there is a bunch of different ways to do it, but that's I think in principle what we are doing here.
Then, you see what things appear to work best, and then you probably do some more tests, and you repeat until you are at a place where you have made a decision on what appears to the best performing way to go. Is that a fair way to summarize the process at a high level?
Tim Ash: Yes, there are a couple of very important things in what you just said. One is that you have to test everything in parallel. I hear about a lot of people doing various kinds of online marketing testing, where they test one version for a week, and then a different version for another week. The problem is there are a lot of things that are different about those two weeks.
Eric Enge: Right. So, if you are in the mortgage market and the FED announces a three quarter percent cut in the interest rate.
Tim Ash: Right. Or, it's a holiday period, or some public relations event happens, and all of a sudden your company gets exposure in the news. Well, that didn't happen the previous week, so you were not really comparing apples to apples. To control for all of those other outside factors, you have to show these different versions of the page in parallel and do it serially as a very last resort.
The other important thing you mentioned is it's important to show all of the different versions randomly to your new visitor. One person shows up, and they see the original. The next person shows up, and they will see something else. They versions are split equally and randomly among all the visitors to your site.
Eric Enge: Right. It strikes me there is one other important thing you have to do is define what you are trying to accomplish with your site and the various web pages in question.
Tim Ash: Yes. If you don't have a clear goal in mind, you are missing the boat. This is about economic value, so you have to decide which parts of your site are mission critical. In other words, your business would grind to a hold if they didn't exist, and that's not your job section, or your news and event section. It's the actionable parts of your site where you try to get people to sign up, download, or buy something.
Once you have identified them you can assign an economic value to each of the actions that you want people to take. That gives you a basis for comparing tradeoffs among all those actions. You may want someone to sign up for an email newsletter, or you may want them to buy something right away. Both of those have a value.
If you emphasize the newsletter, you get a lot more of those sign ups and vice-versa. By defining a dollar revenue value to each of those conversion goals, you can reduce it to a revenue per visitor number that you are trying to optimize for.
Eric Enge: Absolutely. Many people try to optimize their site without actually fully defining its purpose. In the example you just gave you might do some things to increase your newsletter signup, but lower your sales. And, the real objective is the sales. You may you lost sight of where you are trying to go with the whole thing.
Tim Ash: Yes, and, it doesn't have to be a financial goal. Many cases especially for services or products that are complicated or have a long sales cycle or are very expensive, you may have some intermediate goals. For instance, you may want somebody to download a white paper or even spend sometime looking at information on your site.
If you measure for instance the number of people that spend more than two minutes looking at your page, that's a conversion. Or, alternatively you could just measure their time on the page, and try to maximize that. Those are legitimate goals as well, but they have to lead back ultimately to something that has economic value.
Eric Enge: Right. Once you've have some of this basic groundwork in place, how do you select the pages you want to optimize?
Tim Ash: They should jump out at you, it's going to be the mission critical parts of your site. So, you determine the actions you want somebody to take. Let's say it's to buy something; well they land on your site somewhere, and then eventually they can go through your site and end up on your buy page, and then your checkout process whatever that maybe.
You can perform landing page optimization on anything involving that interaction. For instance, you can tweak the actual landing pages on which they arrive, or you can tweak your checkout process which ultimately anybody who is buying will have to go through.
You can open up the top of the funnel, and make the arrival of people more efficient, or you can open up the bottom of the funnel and make sure that more people transact efficiently. And, those are multiplicative effects, so you ultimately want to fix the whole process, but you might start with just one piece of it.
Eric Enge: Right. You might start with one piece, and you might then go to the next piece. But, there is also situation, and you just mentioned one of them where you are dealing with a funnel where you might want to do something that cuts across the entire funnel.
Tim Ash: Yes. Commerce catalogues are all pretty much constructed in the same way. You have a homepage, you have category pages; maybe an internal search results page. Then you have a product detail, and then all of the steps in the checkout process.
There maybe a total of a dozen or so pages on your site. You can optimize individually those pages, or you can say fix the whole shell of your site; the whole page shell of the catalogue to make that more user friendly and easier to navigate, and that's going to affect anybody on any page of your site.
Eric Enge: Right. And then, you can do simple things like change the look and feel of your logo, which probably is on every page of the site, or the global navigation. I don't know how much impact those things will have, but template changes can be made, which are literally affecting every page on the site.
Tim Ash: Exactly. We call that changing the shell of the site.
Eric Enge: For purposes of the rest of our discussion, let's just focus on a single page. Assume we are working on one page at a time, which is the major landing page for a pay per click campaign. After you pick the page, how do you decide what elements to tune?
Tim Ash: If you have a single landing page, there are lots of things you need to think about in terms of choosing what to test; probably the most important is the breadth of impact. So, you want to focus on the most important conversion actions. In that example of buying versus signing up for a newsletter, presumably the purchase is far more important than signing up for a newsletter.
You also want to focus on the biggest possible audience.
If you have several different landing pages, you want to fix the one that is currently running the most economic value through it. We hear a lot of people come to us and ask us to fix a given page, but, we find out it only has 5% of their total traffic or conversions. We always focus them back on the pages that have the most value, even though they often think those pages are already superstars. Because, even if we double the one that's 5% of your traffic, it's not going to have nearly as much impact if we tweak the other one by just 20%, 30%.
You always should focus on the one that has the biggest economic value. That's often parallel to the amount of traffic, so you want to look at the most popular path through your site, and fix that first. Back to ecommerce catalogue example, if your brand isn't well-known, you might have traffic arriving mostly at your homepage.
If it's already well-known, and you have a pay per click campaign, you might be sending people deep into a site and landing them all in the product detail pages for particular products. So, the product detail page becomes more important than the homepage. So, you have to look at where your traffic is landing on your site as well.
Another consideration is granularity of elements that you test. You can make very small tweaks to your site; the color of your call to action button, or just the call to action itself, the text of it; or tweak your sales copy, or you can go revolutionary and build a radically different page.
What you should test depends on how well you have already optimized the page. If it's already performing well, you might just test small tweaks to it. But, you can only get so far making incremental changes; at some point you will have to do a revolutionary fresh look and see if the whole approach you are using is valid or not.
Eric Enge: Right. One of the things that I noted from your book is that, and we have seen this in eye-tracking studies for quite sometime, is that people's attention is very centered on the top left of the pages. Then they drift over to the center. Does that mean that there should be a real attention on what's over in that top left area of your page?
Tim Ash: Well, not exactly. The upper left is usually used as anchoring and orientation. Visitors often don't know anything about you, or they have some context like a link that they clicked on somewhere to get to you. And, they just want to make sure that they have landed in the right place.
The reason that the eye goes to the upper left is that's usually where the brand is established; that's where your logo is, or your tagline. After they know they are in the right place, they have the freedom to look further. But, there is really nothing you can do about the upper left; it's by convention the place where you place that branding and orientation information. As long as you don't violate that convention, you are fine. But, you are right in the sense that in general people read from upper left to lower right, at least in our alphabet system. And so, the eye scans in that direction; so the upper right corner and the lower left corner are usually ignored.
The emphasis of how you present information on your page is very important, and you can control that by the amount of real estate you give something, or the position on the page; and also whether it's above or below the viewable area of the monitor, typically called "above the fold" in newspaper language.
Eric Enge: Right. The top left is where people go to convince themselves they are in the right place. Also, it's long been a pay per click tidbit of wisdom that when you have someone click through an ad and land on a page from your site that you repeat the keyword back to the user. Does it make sense substantively to think about putting the keyword up in that top left area since that's the orientation spot?
Tim Ash: Not necessarily up there, because people don't expect content to be in that upper left, that's usually part of the page header. But, certainly as the title for the content part or the center of the page; as you mentioned there is a lot of academic research that is pretty solid on information forging theory. And, the basic idea as long as people feel they are getting closer to their goal, they will continue to move forward.
If they clicked on an ad that promised them something, had a certain keyword in it; that keyword better be the main topic of the page, so putting in the page title, and putting it permanently elsewhere on the page, probably makes a lot of sense. That way they feel a certain continuity and willingness to invest more of their attention.
Eric Enge: Right. What about tweaking the call to action?
Tim Ash: The call to action maybe the most important part of a single-purpose landing page along with the page headline. When we do testing of the call to action, we get very granular. What I mean by that is we will change the position of the call to action; we will change if it's a button say the background color, we will certainly test different text on the call to action, because just getting exactly the right combination on that call to action is often the biggest driver of conversion.
It may seem to many people that the button color is not going to make a difference. Well, you'd be very surprised, and the other same goes for the headline. If it's just text, it's not very difficult from an implementation standpoint. Changed text on a webpage is often the biggest driver.
We had one case study where we had a company with a lead generation form for their pay-per-click campaign. They changed the headline, and they got a 58% lift in conversion, just by changing the headline. From 'free quote request' to 'instant quote'; now most of us wouldn't think that that's a radically different headline, but it was.
Eric Enge: People don't like the word free. What about the offer itself?
Tim Ash: The call to action has to be consistent with the offer. This moves to a point about language and copywriting, any call to action should complete, or any link on your site, or a button on your site should complete the sentence "I want to …". The person should be able to put their wishes, or their desires, or their task, or what they currently have in mind in there. The call to action might be "download my free white paper", or "buy it now", or "sign up for the newsletter". Those are active, and they can as you can see fit into that "I want to …" formulation. Bad calls to action, are things like "submit", or "click here"; they are meaningless, they don't say anything to the user about what they get if they actually push the button or click on the link.
Eric Enge: Right. The offer is 10 dollars off versus 12 dollars off.
Tim Ash: "I want to get my discount", or "I want to get my 10 dollars off", might be the call to action, and how you tie the two together.
Eric Enge: When does it make sense to do a wholesale redesign as opposed to tweaking just a few things?
Tim Ash: Well, there are two different circumstances in which I'd recommend a wholesale redesign. The first is when you don't have a lot of data; then you really have no choice. You can't do this granular multivariate testing unless you have a lot of transactions. You may have just enough traffic to do a simple A-B split test; so show half of your audience your original page, and half of your audience your new page.
If you have lots of good ideas about the problems of your current page, it would take you too long to do that in a granular fashion. First change the headline, then change the call to action; then change the offer, the graphic, the product graphic that appears on the page, etc. Doing back to back A-B split tests would take too long, so you basically put all of your best thinking into a completely redesigned page and test that against the original.
Then, in one test, you can get significant improvements; you may not be able to figure out which of the changes you made really matters. But, ultimately who cares as long as it performs a lot better, and you don't have to wait for a long time to get that result.
The other scenario is that it depends on how optimized your current page is. If you have already done a lot of testing, then chances are you will only be able to make incremental small changes to it, and maintain the larger framework of how the page is laid out; the emphasis of elements on the page and so on. But, if it's a new concept, and you are just doing this continuous challenger-champion format, then you want to try radical ideas. You want to see if you can beat your champion with your new challengers.
Bringing in outside perspectives by having people outside your company come up with the ideas for the testing can often have huge benefits. Often you can't get that big improvement, because you are the core cause of the problem; you are the person that created the page and it's your baby.
Eric Enge: Right. I interviewed Jonathan Mendez a while back, and his advice was to be prepared to be radical.
Tim Ash: Yes. It doesn't mean you have to blow everything in the test up, because that clearly has its own risk. You may have significantly worse performance in a radically different design than what you already have. And so, the way that you can reconcile those two is to only devote a small fraction of your traffic to radically different version, just to make sure it doesn't really, really stink; and then ramp up the traffic to it so you can finish your test more quickly.
If it becomes clear that it's much worse than you can just stop the test, and cut your losses and try another radical idea.
Eric Enge: If there is the opportunity for large gains, it seems like you'd have a better chance of getting there with fairly radical changes rather than a tweak or two.
Tim Ash: Yes, that's true. Key parts of your page like the headline, the offer, and the call to action can by themselves produce huge improvements. But, yeah you are right, usually if you change the font of the text in your page footer, no one is going to care.
Eric Enge: Right. So, when does it make sense to simply tweak a small number of elements instead of doing whole sale redesigns?
Tim Ash: Well, the nice thing about tweaking small elements especially if they are just text and the copy on the page is it's very easy to do. To convince somebody to do a wholesale redesign of a page that might be a big upfront investment for some companies especially with their QA process, and the approvals and everything else that has to happen.
They may not be willing to take a flier on investing their money upfront and not knowing what the result is going to be. But, if you can show them a small win by changing the headline and getting results, then you can use that as the momentum to build towards more aggressive testing, that's more wholesale in nature.
Eric Enge: Right. So, take it a step at a time, but as you said before this is a lot easier to do when you have a lot of data.
Tim Ash: Yes. That brings up a great point; there are two mainstream approaches out there for testing landing pages. In addition to that, we have our own proprietary approach which I can talk about.
With A-B split testing, if you are just going to test two versions of a page, you need at least ten conversions a day. A lot of people when they talk to us about landing page tests, they say well I have this many visitors to my site, but the number of visitors is really irrelevant. What's important is how many of those conversion actions are happening.
For multivariate testing, and this is again a minimum, I'd say about fifty per day. And then, our own proprietary TuningEngine, which is a non-parametric approach, very different from multivariate testing, you need at least a hundred a day, and preferably many more. With some of the more powerful methods like ours that are available, you can test upwards of ten million versions of a page in a single test.
Most of the time we see people doing small-scale multivariate tests with maybe a few variables on the page, each with two or three different possible versions. So, if you multiply that out, you are talking maybe a few dozen to at most a few hundred different versions of the page in play.
Eric Enge: Indeed. Things can get very complex when you are dealing with that many versions.
Tim Ash: Yes. During my presentations I ask people to guess at the right combination of features on one of the tests that we've done. There are twelve elements in that test; and usually by the time you get to the second or a third one, all the hands in the room have gone down and no one is going to end up with the winning version.
Nobody has guessed right, and the reason is pretty simple; doesn't matter how smart we are as marketers, when you are talking about a test with a million different versions, that's like winning the lottery. I mean nobody is going to be right. You really have to trust the math and the statistics, and collect enough data to know what's the best answer.
Eric Enge: What's your take on audience segmentation and how that fits in?
Tim Ash: The more data you have the better. Typically what you are doing in a test, at least our approach first is to get as much data as you can; as many of your traffic sources that are stable and reliable, and tune for that mix. In other words you are tuning in affect for your average audience member.
If you are focusing on things like usability, and just basically getting out of their way and making their experience smooth; then you can get significant improvements. Beyond that, you may have differences in the psychology or the demographics to your audience. And, they may respond to other factors, for instance your offer.
If you are going to test the offer, then chances are your traffic source is going to significantly affect the result. So, if you have enough data by all means break it down for instance say, your search engine optimized traffic hitting your homepage is different then your pay per click traffic. Break those out and do separate tests on it, but you have to have enough traffic to run the test separately on that subset. And, if you don't, you are often just forced to again just collapse it all into one big pool of traffic.
Eric Enge: Right. And, you talked about this a little bit before. I mean one of the big challenges is to make sure you are getting enough data into the system; and that's when you talked about for A-B testing you'd want to have ten conversions a day, and for multivariate maybe a hundred a day. But, if you are trying to deal with audience segmentation, that's dividing that data down even further.
Tim Ash: Exactly. Yeah, and it makes less of it available for each segment, and it makes your test run longer, and so on. It's an advantage that large companies have that have a lot of conversion action. A conversion action doesn't have to be a sale; it can be like you said a click through to an important page, or a download, or email signup.
Many companies have thousands or tens of thousands of those a day; and that gives them a lot of flexibility. And, they are the ones that can really benefit the most from very powerful large scale testing.
Eric Enge: Right. When you are in a scenario like that, your objective then becomes continuous testing.
Tim Ash: Yes. And, that brings up a really important point how much should you tweak a page; how many tests should you run on it? Well, that depends on how powerful your test is; how many elements you could test at one time, and also how much effort has gone into optimizing the page? Usually, you have your best ideas at the beginning, and you have the most freedom to improve things, because your page is as bad as it's going to get.
Presumably later on, after a few tests it's getting better and better, and it's going to be harder and harder to beat that baseline. At some point you have to just say okay, I am not seeing radical improvements anymore, and I have to move on, and tune another page on our site.
Then, it's a judgment call that a lot of people try to optimize after their page is already optimized; then we basically use the rule of 80/20, get the 20% of improvements on the page that are going to give you 80% of the possible conversion lift and then move on.
Eric Enge: Right. One of the other things that you'd need to remember too I think is that seasonality can be a factor.
Tim Ash: The only two engagements we've had that didn't result in conversion improvements were both in the travel industry. That's not surprising; because one of the assumptions behind the statistics of landing page testing is that your audience doesn't change. So, if you can find what matters to the folks that you sampled and collect the data on in your test; presumably if the same population of people keeps coming to your site, and they will behave the same in the future.
If your audience actually changes, then all bets are off. In travel, you get different people flying different places depending on the time of year; so that's a huge factor. In a lot of businesses, you have very strong seasonal factors, for example leading up to key holiday. People for instance, might buy flowers on mother's day; those that order those flowers a month ahead of time are probably very structured, and value driven, and are going to look for lowest price. Those that buy two days before mother's day don't bother with price.
Eric Enge: They only care that you can get it there.
Tim Ash: Exactly. They care about whether they can be delivered on time, and they will pay a premium for that. The audience will segment itself based on the seasonality of your business. But, if there are different audience segments, they may behave radically differently. It's very hard to tune for spiky short term events.
Eric Enge: Indeed. Another factor here is that, you can't assume that the elements you are tuning, when you tune multiple elements, are truly independent of one another.
Tim Ash: You've just brought out probably the biggest worm in testing and landing page optimization. One of the things that we as marketers try to do is create synergy among all of the elements of our landing page, right? You want your headline to support your sales copy, and you want that to support your call to action.
All of the elements on the page should play nice together. They should be synergistic; the whole is greater than just some of the parts. Unfortunately, A-B split testing which just tests one element at a time, and multivariate testing approaches like the Taguchi Method which is the most commonly used; all assume that the things you test are independent of each other.
The best headline is the best headline regardless of the context, and that's not true at all. It depends on everything else on the page, so it's very dangerous to assume that there are no variable interactions on the page, or dependencies among the things you are testing. That's one reason we developed our TuningEngine technology, because it does take variable interactions into account. We have talked to some folks at Google, and we are one of their Google Website Optimizer Authorized Consultants. They use their own technology to do landing page testing on their own properties.
For instance on the Picasso, picture sharing, signup page, what they found were huge variable interactions. To assume that the things on the page are unconnected in terms of their performance is very dangerous. A multivariate test will give you a prediction of which combination of variables is the best. But then, you actually have to do a follow-up head-to-head test to make sure that that improvement holds up. Because, often if there are variable interactions, you won't get the predicted amount of improvement or anywhere close to it.
Eric Enge: After you have reached some conclusions, you really need to look at that and take another cut at it. You may have tried five different scenarios say, and what might actually be best for you is some combination of elements that you've changed in two and four as opposed to any other things you actually tested.
Tim Ash: What you are referring to is what are called fractional factorial methods like the Taguchi Method, you don't actually sample every possible version of the page.
Eric Enge: Right. No I understand that with your tool, you literally do sample every possible version of the page.
Tim Ash: Right. Goggle Website Optimizer also does that. When it collets the data, it's sampling evenly across all of the versions in play; and that's by far the preferred approach because of these very strong variable interactions. Even if the analysis you do subsequently just looks at the variables individually, you are going to be much better off sampling evenly across all of your variations.
Eric Enge: Right. Even after you are done coming up with the best possible page A that you are willing to do; let's say page A is your landing page. Then, page B is another step into your funnel, and you start tinkering with page B; you may find that things you changed on page A, and things you propose to change on page B also interact.
Tim Ash: Right. Let's say you have two different headlines and two different calls to action. Well, it's possible that if you lock in the wrong headline, you'll never get to the right answer, because of that headline, and the call to action is not the best one that you would have chosen had you had the original headline.
If you do back-to-back tests whether they are A-B split or multivariate tests, it's possible that you will lead yourself away from the best solution and never be able to find your way back to it.
Eric Enge: That makes it fun to think about. What about testing themes?
Tim Ash: We've come up with some testing themes that are pretty powerful. One of the most important ones is that "less is more"; that a lot of times you shouldn't be thinking of alternatives to elements on your page. You should be thinking about cutting them out altogether.
Study after study has shown that if you decrease the amount of information on a webpage, that will increase the time on the page, increase the retention of the information and increase your conversion rate.
For a lot of sites, un-cluttering your page is the most powerful thing you can do. You have to ask yourself how radically can I simplify this page. What do people actually care about given they've devoted very little attention and time to this page. Just like the billboard on a freeway; you are driving by 65 miles, 75 miles an hour, and the rule of thumb with billboards is it can't have more then seven words. Well, the same discipline should be applied to landing pages; every word should be scrutinized for its importance and every visual to make the pages uncluttered and clean and as inviting as possible.
"Personalizing" it of course is also a powerful theme, and "testing the offer" as we have already covered is critical. Personalization would be something like following up with the keyword that they typed in as the page title or using other information that you have about the person on previous visits to your site to personalize it further.
Eric Enge: Right, and if you know their name for example.
Tim Ash: Exactly. And then, on testing the offer, the offer itself, the total solution, the headline, the images you choose, the call to action text and format, limited availability, and those kinds of things are all aspects of the offer; and of course the price if you are trying to get them to buy something.
Eric Enge: Alright, thanks a lot Tim.
Tim Ash: Thanks Eric, I really enjoyed chatting with you!.
Have comments or want to discuss? You can comment on the Tim Ash interview here.
Other Recent Interviews
- Matt McGowan - March 4, 2008
- Avinash Kaushik - February 18, 2008
- Danny Sullivan - February 11, 2007
- Google's Adam Lasnik - February 4, 2008
- comScore's James Lamberti - January 28, 2008
- Incisive's Kevin Ryan - January 28, 2008
- Eurekster's Grant Ryan - January 14, 2008
- Eurekster's Steven Marder - January 7, 2008
- Google's Sep Kamvar - December 17, 2007
- Microsoft's Grad Conn - December 10, 2007
- Seth Godin - November 26, 2007
- Microsoft's Ramez Naam - November 12, 2007
- Google's Matt Cutts - October 8, 2007
About the Author
Eric Enge is the Founder and President of Stone Temple Consulting (STC). STC offers Internet marketing optimization services, including SEO, Social Media and PPC optimization, and its web site can be found at: http://www.stonetemple.com.