Internet Marketing with Mike Moran and Lee Odden, Part 5 of 8 (Audio Podcast Transcript)
- Apr 13, 2009
Editor's Note: This is episode 5 of 8. If you are just jumping in, you might want to start with the transcript of episode 1.
Lee Odden: Welcome to IBM Press podcast series with Mike Moran and Lee Odden. I’m Lee Odden, CEO of TopRank Online Marketing and Executive Editor of http://www.marketingblog.com/". TopRank is an internet marketing consulting agency that provides enterprise search engine optimization, social media, and online public relation services for clients that range from Hewlett-Packard to McKesson. Our guest is Mike Moran, distinguished engineer with IBM. The credentials list on Mike Moran is nearly a mile long including the fact that he is an author of two very important books, the new http://www.informit.com/store/product.aspx?isbn=0132255960" and http://www.informit.com/store/product.aspx?isbn=0136068685" Mike has more than 20 years experience in search technology and he led the original search marketing strategy for IBM.com. He’s also been responsible for the integration of IBM site search technologies. Mike worked on IBM's website for eight years and now works on IBM's OmniFind Enterprise Search and Analytics Products. In addition to his search work, Mike is also a columnist for Revenue Magazine and WebProNews. He also writes a very popular blog called “Biznology,” which you can find at http://www.mikemoran.com/". We’re in our fifth podcast of our series of eight from IBM Press, and so we pose the question about the book — it’s called Do It Wrong Quickly — how do you go about doing that, how do you start, how do you know what wrong is? That's really the overall topic of this particular episode in our series.
LO: So, Mike, could you break it down — Do It Wrong Quickly? How do you know what it is? Where do you start?
Mike Moran: I think a lot of it depends on where you are now; there is no ‘only answer’ to tell a company, “Hey, start here.” What you can do, however, is to look at whatever intractable problems you seem to have and look at the organization and find out what are the things that people are worried about. What are the things that they are always discussing? What are the things that they never seem to be able to get right? So maybe the organization says, “We can't seem to get our e-mails opened. We don't know why they open rate seems to go down every month.” Now is not the time to explain to them, “Well, we shouldn't really be focusing exclusively on open rate. You know, do you care if anybody bought anything based on those e-mails?” It's not the time to evangelize and tell them what the right metrics are. It's the time to focus on helping them understand that whatever their problem is — as long as they can choose a metric — that’s the one that they want to be judged by. So if it is open rate, then what you do is say, “Well, okay, you guys have tried 15 different things to raise your open rate, but how about we just admit we don't know how to raise our open rate? The real goal for us is to experiment. The real goal is for us is to say let's try 15 different titles for the next e-mail that goes out, and let's see which one of them gets the highest open rate, and then we use that for the entire campaign after we run our test.” That's the way direct marketers work. If you can pick a really difficult problem and get the organization to buy-in to the fact that we’ve tried five different things, we’ve had task forces, we’ve had all the meetings, and nobody seems to be able to come up with the magic bullet. Instead, what we want to do is we want to try as many different things in fast as the time as possible. That's the place you start because what happens is when an organization is galvanized by a problem that it just can't solve, then it's willing to try new things. They won't fall back on the old ways because they aren’t working. So the “it” for Do It Wrong Quickly is the place to start; it’s with a really difficult problem that people haven't been able to figure out, that has a metric associated with it, and that gives you the feedback to tell you whether it's getting better or not.
LO: Well, that sounds like a key that there is a metric associated with it. It's a measurable outcome. There is a foreseeable goal that can be achieved. And it also sounds like it's really about implementing a process for testing, experimentation, and willingness in the organization to follow those sorts of problem solving techniques. I understand using wrong is just an expression, but how do you know how wrong it is?
MM: I think it goes back to knowing what metric you are going to use to measure it. So if you’ve got a system that can tell you how things are changing, it's almost not important whether the metric is accurate or not — and some people are aghast when I say this — but you can have a system that really isn't doing a tremendous job at counting things properly. You can have all the metrics experts come in and tell you, “Well, we don't really know that that was how many page views we had, or we can't actually count how many e-mails were opened because if people aren't connected to the internet when they open the e-mail we can't count.” They’ll give you all kinds of technical explanations for why it isn't totally accurate, but what is really important is not whether that exact number is accurate — what's really important is the trend. What's important is the difference between the number we checked for scenario A and the number we got for scenario B. They are both probably somewhat inaccurate, but they are both probably inaccurate in the same ways because we’ve used the same method to test them both, and the difference between those two things is something you can really count on. The difference between those things tell you if it's getting better or it's getting worse. So that's what you really have to focus on. So there are a lot of folks who might get caught up in the idea of whether something is accurate or not, or whether it's exact, or whether it's precise, and you don't need that. I mean, what you really need is to say the trend is my friend. So whatever it is that the numbers are showing you over time — that's what helps you make decisions about what to do next. So that's how you know how wrong you are, and it's how you know each time you change it whether you are making it a little less wrong everyday.
LO: So if your testing methodology is consistent, then at least you can rely on the relative difference between one test and another?
MM: That's right.
LO: So when you say you “Do It Wrong Quickly,” at least in my experience, and I am sharing yours too, most companies are challenged at doing anything quickly. So what advice do you have on how companies can speed up the corporate metabolism?
MM: I think that one of the things that you have to look at is what it is that you are trying to do. I think that sometimes we try and sub-optimize things. We try and say, “Okay, well, how can we do the content the fastest, or how can we save the most money on the content, and you shouldn't be doing things like that.” We should be looking at it as how do we get to a better experience where we’re getting higher conversions the fastest. That might mean doubling or tripling the amount of money that you spend on content because you’re going to create eight different versions of the e-mail before you send out the one to the entire mailing list. Because you’re going to test those eight, you’re only going to send out the one that tests the best. And so one of the things to focus on first is what does quickly mean. It means the quickest way for you to increase your conversion rates — the quickest way for you to get more sales. There are lots of different ways that you can do that. You can use something called multivariate testing, which is a big mouthful of words, but what it really means is that the computer can actually check all those eight different versions of the e-mail, and it can tell you which combination of all the different elements of those e-mails is actually the one that would perform the best. It can check millions of different combinations. That's the kind of thing that can order me how fast you get responses because what a lot of people will say is, “That’s nice that you want us to do all this stuff, but who is going to control those numbers?” Well, the computer can control the numbers, and it can tell you this is the one to use.
And the other thing that people asked me a lot — I think especially because I am a distinguished engineer and I work for IBM — is they say, “How is it that you think that things could be done quickly? The technology people that I work with are the slowest people ever.” They are the ones who want a book full of requirements, and they’re going to spend three months writing it all down, and then I have to read the book and tell them what's wrong, and then they’re going to take a year to do things. There is a new technique in software development called agile development and what it basically does it says you have to have software that works all the time. Business people are allowed to say what their requirements are. You want to have a working system every week or two weeks or however long you want to make your periods, so you end up making very small changes very quickly. You don't do these huge projects. We’re going to do this entire site redesign — it takes a year. What you do instead is, you change this page to act like this instead of like this, and then you test to see how well it's working.
So by using those kinds of techniques where you’re accepting the fact that you are going to have higher content costs, you are going to lower your testing cost by automating things like multivariate testing. You’re going to lower your technology costs and really increase the speed and the responsiveness of the technology organization to the business people. Then you’re going to be in a position where you can really not only do things faster, but you probably won't even spend any more money doing anything you do now — you’re just really doing it better.
LO: That's fantastic. And again, great advice and great insight.
LO: That concludes episode five of our set of eight Podcasts with Mike Moran and IBM Press. Somehow this really makes a lot of sense intellectually, but I am sure a lot of our listeners are wondering how marketers can overcome the personal and organizational barriers to change. Well, that's the next topic in episode six of our podcast series. For more information on Mike Moran as well as the two books he has written, http://www.informit.com/store/product.aspx?isbn=0132255960" and http://www.informit.com/store/product.aspx?isbn=0136068685", please visit http://www.mikemoran.com/". This series is brought to you by IBM Press at www.ibmpressbooks.com".