Kevin Lee Didit Co-Founder & Executive Chairman, has been an acknowledged Search Engine Marketing expert since 1995. Kevin’s years of SEM expertise provide the foundation for Didit’s proprietary Maestro search campaign technology. Didit’s unparalleled results and client growth have earned Didit recognition not only among marketers but also as part of the 2007 Inc 500 (#137) as well as a #12 position on Deloitte’s Fast 500.
Kevin’s “Paid Search Strategies” column for ClickZ is read by thousands, and his book, “The Eyes Have It: How to Market in an Age of Divergent Consumers, Media Chaos and Advertising Anarchy” has been widely praised. A founding board member of SEMPO and its first elected Chairman Kevin is also active on DMA and IAB Committees.
The Wall St. Journal, Business Week, The New York Times, Bloomberg, CNET, USA Today, San Jose Mercury News and other press quote Kevin regularly. Kevin lectures at leading industry conferences plus: NYU, Columbia, Fordham and Pace Universities. Kevin’s expertise is also valued by Wall Street he has briefed analysts and clients of JP Morgan, RBC, UBS, Piper Jaffray, Bear Sterns, Citicorp & others. Kevin earned his MBA from Yale School of Management in 1992 and lives in Manhattan with his wife and daughter.
Eric Enge: I would like to start by talking about some of the big events of the past year, and there have been many. You have highlighted some of those in your writing. Let’s start by talking a little bit about the impact of the Google DoubleClick acquisition.
Kevin Lee: Right, which is still pending I believe based on the EU still reviewing the details of the way that acquisition would end up occurring (editor’s note: the EU has since approved the acquisition so it is now complete).
Clearly, there will be some new inventory opportunities that are likely to show up within a certain period after the acquisition within the Adwords Interface. Probably it would be fairly similar to when you opt into contextual advertising, and then the display portion, and/or video portions of that tab.
Google will hypothetically be able to get some converts of fairly major publishers to give them some inventory so that with Doubleclicks assets and joint innovation they would launch a high end advertising exchange.
Such an exchange opens up non-search inventory opportunities for search marketers. The real question is will that combined entity become a really important dashboard for internet marketers as a whole; not just purely search marketers, but internet marketers as a whole who are trying to do some combination of branding, direct response, or other integrated marketing.
Eric Enge: Right. One of the things that you have spoken about is the blurring of the lines.
Kevin Lee: Yeah. I mean the investment community still loves to think about internet advertising as being strictly search or display, when in fact that’s just what the consumer sees, either something textural in nature, or something graphical in nature. When in fact those of us who use it everyday know that search is really more of what the consumer does that resulted in them seeing an ad of some kind, graphical or otherwise. The categories of search vs. display have been there since Goto.com, where generally what people saw as a result of a search was group of text links, either paid links or organic links.
But, for a while there in the early days if Internet marketing, you could actually buy keyword based banners. So, there is no rule that says you can’t have banner advertising, video advertising, or rich media advertising that is targeted based on real time search behavior. Or, of course with the advent of behavioral targeting technology, that form of advertising could even be cookie-based targeting of individuals post search, based on information they provided, or needs that they expressed as a result of their search behavior.
Eric Enge: Right. It’s interesting that people want to silo the two advertising models into two different categories. But, at the end of the day the objectives of advertisers that are buying those ads are the same.
Kevin Lee: Yes, absolutely. Although advertisers sometimes have somewhat irrational silos, the whole idea behind having one budget for branding and another one for direct response, as if the advertising you are doing with those separate budgets contributes nothing to the other side.
When in fact clearly there is a direct response aspect to brand advertising and direct response advertising results in brand lift. Both at the ad level and even more so based on the engagements that occur after a person lands on a website has a chance to surf round for anything from a few seconds to hours potentially.
Clearly that post-ad behavior on the site probably results in the change in the branding metric. So, there is an integrated view or holistic view of marketing where one starts to think less about whether objectives are purely direct response or purely branding, and more about wanting to move more consumers closer to making a purchase from the advertiser through advertising and the web-site.
If the consumer moves from the early stage of the buying cycle or the buying funnel to the second stage; or they move from the last stage into actually making the purchase online with a trackable cookie, those could both be good, right?
Eric Enge: Yes.
Kevin Lee: As long as they move that consumer closer to a positive purchase decision, and exert an influence of that consumer, it’s a win. Of course measurability of the earlier stage influence is a lot more challenging.
Eric Enge: That’s right. It turns out to be a very complex problem; we will talk about that a little bit more later. What about the acquisitions by Microsoft and Yahoo? What are your thoughts on those?
Kevin Lee: Sure. Of course now Microsoft is trying to acquire Yahoo, but they have each individually made a lot of acquisitions themselves. The Microsoft acquisition of Aquantive is fairly similar to the Google DoubleClick deal, there are various elements of the acquisition that are interesting.
One big element is the third-party ad servers and the publisher side ad server division which has contact with the publishing community. The advertiser side ad server is something that marketers are already using to track conversions. So, the irony of course in that is up until fairly recently if you were to ask many marketing and advertising folks in the internet space to name some third party ad servers, they’d probably say Atlas, Dart, and Mediaplex, regardless of the fact that those will no longer be third party ad servers. Clearly, Atlas is not really truly a third party ad server anymore; now that one of the largest internet publishers owns it.
Similarly, once the DoubleClick deal closes; that won’t be a third party ad server either. Hypothetically the Chinese Walls would be erected so that marketers can still confidently use those platforms to track their performance that they are buying from the same exact company that’s tracking in. That’s a real shift in the internet marketing landscape, but when I talked to old-school media people, they just scratched their heads and they said I can’t believe that. If GE, which owns NBC, bought Neilson things in the TV ad-buying world would erupt, and nobody would buy anything.
Eric Enge: It’s the new age of marketing I guess, right?
Kevin Lee: I guess so.
Eric Enge: Not having real walls in some of these cases can be scary.
Kevin Lee: Some of the more traditional marketers and advertisers have been around for decades. They are a little concerned about that, and really want to be reassured that if they are going to continue to use technology owned by a network or a publisher that they can trust it either to provide the right data or n some cases where it is making bid changes to make the right decisions, and that there is no bias either direction, right?
Eric Enge: That’s right.
Kevin Lee: So, as far as other similarities between acquisitions; Yahoo having purchased Right Media which is an ad exchange, and Microsoft’s acquisition of AdECN. The Right Media has dramatic size and scale over AdECN, but they are very similar businesses.
They share the idea that an ad network can be more than just an ad network; it can actually be a network of networks. The advertiser eventually gets served a particular impression, particularly in display media. It could actually come through a fairly convoluted set of hops that goes through an exchange based on that exchange acting as an intermediary. Right Media proved that there is value in that now; a lot of ad networks are actually trying to reposition themselves as exchanges, because apparently that’s what corporations want to buy or acquire.
Eric Enge: Gosh, offering people what they want to buy; I guess that sounds like a pretty interesting business plan.
Kevin Lee: It’s the flavor of the month to some extent, because two big players have already made their purchases. So where will the five or six other ad exchanges sell to if not Google. Google bought DoubleClick, and DoubleClick’s got its media exchange already, they call it the Advertising Exchange.
So, the big sales have already happened. You are going to have to actually build a viable business, and not just build something that you think somebody wants to buy at a corporate level.
Eric Enge: Right. What about Yahoo’s Panama and shifts in Google’s Quality Scoring Algorithm?
Kevin Lee: Sure. It was actually surprising how short lived the impact of Panama was at least for our advertisers; that you either got bit of a bonus or benefit based on the algorithmic change, or, you got slap as having a low predicted click through rate in comparison to your neighbors in Panama. That shook its way out within a couple of months, maybe three months, and then after that advertiser were back in equilibrium wherever they ended up.
You simply had to use whatever best practices were appropriate if you got negatively impacted; obviously you have other levers at your disposal; the easy one is to raise bid price. That’s also the most expensive one, so there are other levers that you might want to use to convince the Yahoo system that you are in fact more relevant and are in fact going to get a higher click through rate than your neighbors, all other things equal.
Similarly there have been a lot of changes, and I’d call them more refinements to Google’s Paid Search Algorithm over the last couple of years. And, they continue to make their quality score algorithm a blacker and blacker box. They have made the algorithm much more opaque by throwing more and more variables into it including things like match of the landing page to the keyword that was being advertised, which they’ve told me is more of a binary factor than a continuum.
You are either a good fit or a bad fit; it’s not really a gradient. But, nonetheless that’s an external factor that didn’t need to be taken into account before; and not good for those advertisers who built Flash-based microsites without any copy at all to try to land people on them for a rich-media experience.
Eric Enge: They were by definition not a good page.
Kevin Lee: Some people say that they end up giving a neutral rating, which wouldn’t potentially count against you. But, even the Google team can’t necessarily agree on exactly how those On-page factors might influence one’s quality score. Then, all the other factors that end up getting thrown in to that soup, make it increasingly difficult for a marketer to predict ahead of time what the cause and affect will be of any campaign changes or one landing page over another.
Eric Enge: Indeed. It’s amazing how different the world is from just even two years ago.
Kevin Lee: Absolutely. We used to have a completely transparent bid landscape, and you knew exactly what the cause and affect was of a bid change. Now, whatever technology you are going to be using is going to have to be able to essentially test the elasticity of the market with regards to bid price changes, or changes in reserve price on the bid or max bid and the resulting changes in position.
Whether one wants to test those things on a national basis or even on a geo-segmented basis, it gets pretty complicated. It pretty difficult arithmetic; and for any large scale market it’s more and more difficult to manage that in a spreadsheet with a pivot table. It’s going to be something that either the engines try to solve, which of course they are ones that might mess it up in the first place, or a third party technology provider like Didit and our competition will continue to solve that challenge.
Eric Enge: Yes. Let’s switch directions and talk about what’s happening in terms of search getting integrated with other media. A few months back I published an article related to some research done at Range Online Media. That article showed that display advertising campaigns had a material affect on click through rates on pay per click campaigns. And, in fact what they were recommending is that if your big season is Christmas; maybe in September, October you should ramp up your display advertising and ramp down your lower margin keywords. Then, as you get closer to the hot season, ramp down your display and ramp up your spend in pay per click.
Kevin Lee: Right. The earliest I can recall is actually with Harrisdirect, before they got acquired by E*Trade in 2006. Harris was a client at the time and we and Yahoo saw definite interaction affect between major online ad spending, and media, and search behavior. What we’ve validated since then is that it’s even much more of a macro phenomenon. We have asked our advertisers when possible to share with us all the media, all the PR, and all the direct mail, or direct response advertising that they are doing; and just give us a schedule of when things are happening.
We can see what the impact of that is on search behavior. And, it correlates pretty well; it actually correlates so well that to some extent search ends up being the metric of the success. For example, if you were doing a good job buying that media or/and a good job with the creative of that media, you would expect the level of search to go up more; and if you were doing a horrible job either buying that media or with the creative of that media, you’d expect less of an impact.
comScore does a final study every year after the Super Bowl that shows the increase in traffic to sites that were advertised during the Super Bowl versus the Sunday prior. It gives you an interesting illustration of that at work in a very siloed environment.
We look at a client’s media spending; it ends up being important for us predictably, because we could set our technology up to be more attuned to any changes that might be occurring. They happen most broadly on brand keywords, but they even impact other keywords as well.
Eric Enge: Then, you get some of the reverse too, right? You get pay per click campaigns, where people go and they do research, and they want your products; but then they want to see it before they buy it, so they go to the store.
Kevin Lee: Yes. There is a whole offline and online thing that ends up being really important for multi-channel retailers. Obviously, it’s not important at all for Amazon; as a matter of fact they may curse it, because they don’t have a retail channel. The last thing they want is you finding the book on Amazon, and going to a local Barnes & Noble to buy it. Or, finding a flat screen TV on Amazon and going to the local Circuit City or Best Buy to purchase even if it’s a couple of bucks more, and you want to touch it, see it and perhaps ask a salesperson.
There is still a lot of leaky data in the system, right? I mean we think of search and online advertising as being this great feedback loop system, where we can spend our money, and we can see exactly where it’s working. Sure, we can see better than we could ever see before, but there is still a lot of areas for data leakage; and taking into account that leakage of data ends up being pretty important in running campaigns in an effective manner.
If you might look at different segments of a campaign and determine that the online to offline leakage is significantly more within certain segments of your campaign then other. I’ll use a fictitious example; if Circuit City which is not a client, but if they sell DVDs, MP3 players, and they also sell flat screen TVs, they’ll see different behaviors. The MP3 player and DVDs may be more of an impulse sale much more likely to occur online, and the flat screen TV due to shipping cost and breakage might be something where people are more comfortable buying them in a store. Figuring out how to build models around those situations ends up being really important for the multi-channel merchant.
Eric Enge: Right. Best case you are going to have a lot of estimating going on.
Kevin: Yes. You just want to make the estimate directionally accurate, and as accurate as possible. Some of the stuff you can actually just think about, and know at least which direction to put in, i.e. you know plasma TV is probably more likely to be bought in the store. But, if you can validate that information, and get a magnitude of the effect, that’s really useful.
Obviously, some retailers have been smart enough to create one option within their shopping cart that says Pick Up at Store, right? So, then they have actually closed the loop online, so there is some online action that translates into an offline behavior.
Eric Enge: Right. Treating search as a strict direct response model is often not the right thing to do anymore, is it?
Kevin Lee: I don’t think so, but I have no illusions that marketers are going to sit bolt upright, and decide they have been doing it all wrong for the last couple of years, and realize that they really need to think more holistically. It’s really endless case studies and endless data that’s needed to convince them, and sometimes it helps if competition whips their butt a little bit, before they finally wake up and realize that they are doing something wrong. The CMO or VP might suddenly question why the competition are always outbidding them, and recognize what are they doing; and why are they bidding so high.
Obviously, a competitive advertiser could be using other data to justify that bid. So, it’s going to require an evolution by marketers into a new way of thinking, and concurrently with that I think the tools will make it easier and easier to estimate accurately what that cross channel affect is; what the number of cookies that are lost based on shopping at work, and buying at home and anti-spyware programs and all the other data leakages that occur.
More importantly, how does search influence people towards a future purchase? That could be an online purchase, an offline purchase; and there could be a lag time of several months in a high involvement purchase.
Eric Enge: Right. You are getting into other kinds of things now that search marketers don’t typically talk about like Reach and Recall, and things that are just the normal language of traditional marketers.
Kevin Lee: Sure. At SES, the panel I am speaking at is a bunch of old timers talking about how search may actually improve the health of your brand. And, how that impact is really being dismissed as gravy, and marketers’ are not giving it any credit.
Eric Enge: What should a search marketer that’s sitting in a silo or someone like you or I, who are focused on the search side of things; how can we help the process in terms of getting traditional marketers to understand what we bring to the table better.
Kevin Lee: I think it’s an educational challenge all the way through the value chain. Usually that education needs to occur from the top down; so the CMO, CEO level. They have the ability to turn the dials to change media mix models a little bit, or at least make sure that there is an impetus towards measuring those interaction affects or understanding them. Often, some of the existing players in that ecosystem are not necessarily have an incentive to shine a flash light on search as more than just direct response.
If you think about the economic model of a lot of these mega agencies and mega agency holding companies; it’s not unusual for them to have negotiated an umbrella percentage of media spend under 10% to manage all of the media of a particular marketer or a client for them.
Certainly when it comes to buying Super Bowl ads, some are more than happy to do that for 6% of spend at two million a pop, right?
Eric Enge: Right.
Kevin Lee: It’s not a lot of work, I can do that. I can make decent money doing that, but it is really hard obviously when you are delving into narrow cast media, which requires lots of little channels working simultaneously. Now, lots of keywords work simultaneously in lots of different segments; that tends to be labor intensive even with the best technology. There is some stuff that still humans do better, so, there is a certain amount of labor involved in that. Somebody needs to pay for that, either it’s an in-house person, it’s an agency person, and I just don’t think a marketer can get is done with 5%, 6%, 7% of spend regardless of where the resources sit.
Unfortunately if a mega agency holding company says people use DVRs more than TVs now; one probably should spend less money on TV, should probably put more of it into search, and less into TV because one can develop landing pages which are contain Rich Media and video, and so wonderful in promoting the brand.
We should send people there. Doing that will cost more man hours per dollar of media spent. That isn’t the kind of media buying that the agency used in calculating their profit margins. So, they talk a good game about being excited about search and rich media online site experiences and ads, but many of them when it comes right down to it are really not that interested in getting more complex with narrow cast media buys replacing broadcast tonnage.
Eric Enge: Right. And, it seems to me as one of the issues is that instead of broadcasting to a blind audience, with search you are actually putting yourselves in front of an audience that’s declared intent.
Kevin Lee: Yes. That intent could have come from a variety of places. It could have come from Buzz Marketing PR; advertising that you were doing as we were talking about earlier. But, the consumer, the searcher did raise their hand, right? They did raise their hand as being in the market, and that’s unique to search; but the real question is somebody who raised their hand and said they are interested in cell phone, the same person who was the first to raise their hand and say they are interested in the Nokia 5310.
Those need to be treated differently as marketing channels. And, that’s also very different from the other kinds of media; with that kind of niche aspect of it. Most media is intrusive; it intrudes upon your day, you are trying to read that newspaper, read that magazine, watch that TV show, and it tries to convince you to think about something completely different.
Search is really the opposite; you wanted to learn about used boat propellers, and so you type that in. Now you are seeing an amazing amount of information relating to used boat propellers with an E-bay listing right up there at the top.
Eric Enge: Right. I think another thing that plays into all of this is just really getting a better understanding of the customer buying behavior and the cycle things they go through; and how all these diverse media opportunities hit you in different places in the cycle.
Kevin Lee: Yes. There is a field they call media mix modeling. It is quite complex; there are some proponents of that are trying to solve it; various PhDs in data modeling. It’s great that at least at both in academic and a business level there are folks attacking that, because it’s going to be necessary in the future particularly as the consumer spends a dramatically increasing percentage of time online. Studying the interaction affects at least within the online portion, where trackability is better is going to be key.
It goes much beyond the concept of the view-through, which was the first attempt at understanding whether or not a particular media exposure had resulted in a site visit. Recently, I actually wrote a column on that. You just have to be very careful in a way to use things like a view-through. But, the concept was rational, and that how do we understand whether or not media is impacting behavior. If it’s impacting attitude, that’s even more difficult to measure.
Then, you have to get into the sampling models and things; things that some of the research firms do.
Eric Enge: It’s a very complex bag of stuff. So, segmentation is a part of this too. A very simple example is someone types in digital camera reviews versus a specific digital camera model number. One of those people is in research mode and another is in probably closer to buy mode. And, they are looking for different experiences once they get to your website.
Kevin Lee: Sure. Segmentation by keyword is just one of the levers of segmentation that we have. Obviously the audience changes as far as who is sitting in front of that browser based on time of day, day of week or geography. So, you’ve got an even more complex segmentation problem with regards to truly understanding, how do I give this person the best user experience?
Locally, it’s all driven by data; so whether you are trying to decide what to do with the people who have typed in early versus mid versus late stage buying keywords, or you are trying to decide how to take a single keyword, typically a power keyword, one with high opportunity and create sub-segments from that. Data is your friend, but it can be quite overwhelming as well.
Eric Enge: Right. Not everybody has the kind of web application necessary to handle this dynamically. So, many have to handle it through static means.
Kevin Lee: Yes. Because of that we actually have ended up building a routing function into Didit’s Maestro so that if a client has different static web pages for different audiences we could actually split up the traffic from a single keyword listing, and send it to different places based on certain parameters.
Eric Enge: Right. That sounds like it heading towards doing something similar to a landing page optimization tool.
Kevin Lee: It is, but the goal is to create the right experience for the user, right? If you’ve got a marketer that has an existing site with a particular laser printer; a category of laser printers, and then also a category of printers. For the same keyword, the right landing page may actually differ based on geography or time of day.
That’s before it starts stripping apart a particular page into it’s component parts like image, background, color, bullet pointed versus paragraph text; all the stuff that an Optimost or an Offermatica or somebody like that would do. Sometimes you just want to test pages you already have against each other for different segments and understand whether or not they work better or worse, for a particular type of click.
Eric Enge: At the end of all this is search going to begin to capture a larger percentage of company budgets?
Kevin Lee: I think so. Obviously there is a bit of a concern right now economically with regards to potential recession coming around the corner. It will be interesting to see how marketers deal with that from a budget allocation perspective, but search is as you have mentioned earlier being managed on a direct response basis now. And, it pays out from that perspective, so I can’t see there being a tremendous cut in search budgets.
There maybe a reduction in supply, because less people maybe shopping. That might result in there being slightly less spending, but it certainly won’t be an arbitrary change for most marketers. As more and more marketers start thinking about search more holistically; start getting the data to understand that those search clicks were actually worth more than they thought they were, I think it’s inevitable that search spending will go up. Of course, part of this interesting fact is that many search marketers define search as anything keyword targeted, right?
So, contextual inventory for example will work for some marketers particularly if they are not purely direct response and are looking for Reach. Contextual ends up in Google or other search budgets, and therefore they think of that as search. You and I may not really think of contextual inventory as search. It’s keyword targeted, but there wasn’t anybody who typed anything into a box, So, that’s not really search; and those budgets will increase even if the pure search budgets don’t increase.
Then you’ve got the whole onset of behavioral retargeting particularly at Microsoft and Yahoo, who are working on rolling stuff out, along those lines. That could open up a whole new era of search; it is targeted display media based on search behavior.
Eric Enge: Right. All this gets again back into the blurring of lines in almost every conceivable direction, right?
Kevin Lee: Yes. If its keyword targeted they think of it as search even though it wasn’t, but it’s still related to search though. It’s going to be very interesting to see how the engines in particular roll that out in the future.
Eric Enge: Right. It’s almost the opposite of blurring of the lines. They are using labels much more broadly then we might have originally thought.
Kevin Lee: Right. I haven’t had a chance to really dive into the meat of the sample of annual SEMPO survey results yet. We are releasing them during SES New York. It will be interesting thing to see how those are different from last year with respect to what both of the agencies and marketers are planning to do.
We’ve got a much higher response than last year; we left the survey open longer to try to continue to get a higher response, and we got over seven hundred. I think last year we were a little under five hundred if I am remembering correctly. So, there is going to be some really good data in there from both the agency and marketer perspective. (Editor: Here are the SEMPO survey results)
Eric Enge: Exactly. Let’s switch gears into our third topic area which is measurement. There are definitely some limitations on analytics tools. You used a good phrase before which applies to analytics as well, which is making sure that you look to the tool to see directional accuracy, so as opposed to absolute raw numbers.
As long as we are directionally accurate; a keyword, or engine, or what time of day that is great vs one that performs poorly and continues to under-perform, in comparison to everything else, we are OK. If you are using the same platform to measure, it doesn’t really matter which platform you are using as long as it is fairly accurate. The different platforms and analytics may disagree. One may say you’ve got a hundred orders; the other one says you’ve got ninety. But, when it drops to fifty, the other one also drops down to forty-five, right?
Eric Enge: Right.
Kevin Lee: So, as long as you are able to do a comparative analysis between several different media opportunities, it’s not a matter of getting the accuracy perfect. I think that’s something that it takes some marketers a long time to digest, because they are thinking that web analytics should be like an accounting package.
Or, your checkbook, where you really want to get it balanced to the penny at the end of the month. Of course my checkbook probably doesn’t balance to that level. Anyway, but you certainly want to get it somewhere in the ball park, but in marketing it’s all about making the best decisions directionally. Many packages will allow you to do that.
Eric Enge: Right. We did an analytics survey basically showed that the analytics package reporting the most traffic could be reporting as much as 50% more than the lowest reporting package. That’s a really wide range, but you’ve got to put this into perspective.
Compare that to a guy who just ran a TV media campaign, and told the viewers to use a special code with their purchase so they can try to track how successful the campaign is. Maybe 20% of the people who actually do act on that TV campaign actually use that code. So, they are getting 20% data and making decisions. They would die for 67% of the data, right?
Kevin Lee: Right. They also typically don’t have the luxury of running five different TV campaigns with five completely different offers so that they can test the relative value of the campaign and the offer simultaneously. We have that in search, and as long as it’s directionally accurate, we can say that the Blackberry cell phone is working better than Nokia’s cell phone by 50%. Even if our raw numbers are off, the fact that one is working better than the other by 50% ends up being an important number.
Obviously, you try to get things as accurate as possible. Generally we go through a reconciliation process with new clients before we even turn on our automated technology. We just do a tracking phase just so that after two weeks we can look at their numbers; look at our numbers and find out exactly what kind of variance we are looking at. That allows us to determine if we have to use some kind of an adjustment factor.
Eric Enge: Right. That type of synchronization makes sense. It’s really about making sure you can speak the same language, right?
Kevin Lee: Sure. It’s about making sure that when we put goals in place, right if the two systems have some kind of a variance, we just want to know what it is. The reason we try to keep the baseline long enough is because we want to make sure that our system has an opportunity to start understanding conversion behavior, right? Some purchases don’t get made in the same session or the same day, and so we want to make look at their system which obviously has been running before we started, and understand how that’s trending.
Eric Enge: Right. What about bid management tools?
Kevin Lee: There are flavors of bid management tools out there including some that are more targeted towards the small size or mid size marketer that are fairly new. I think Clickable is just coming out of beta right now, whereas folks like us have been around for a long time. Some of our competitors have already been acquired a long time ago. Some, like GoToast were acquired by aQuantive long before; aQuantive got acquired by Microsoft.
The real key is trying to figure out what your process is, and what the fit is between the technology, and your process, and your goals. Because, there are lots of flavors of bid management technology, and even the engines have rolled out fairly rudimentary bid management technology themselves. If you are willing to use their tracking pixel, they are willing to change bids for you to hit a particular CPA.
As a result, in its most simplistic form, bid management has been commoditized. I think it’s really a matter of what else comes with that, and what are the different ways in which one can manage a bid? What are the different formula options; what are the different algorithms that one can use to manage bids? Do you manage them where each keyword is its own business; does it have the ability to set it up as a portfolio for marketers who’d prefer to do that?
There is also the aggressiveness level of the bid management technology. All sorts of fun nuances; but not to get too much into the technical issues of the bid management technology. It’s really just more matter of talking to the vendor, and getting a sense of what kinds of businesses that vendor has been successful with before, and seeing whether that’s a similar kind of business, and expressing your business needs to the vendor so that they can give you the kind of information that you will need to make a good decision.
Eric Enge: The early bid management packages offered some very simple methods, and it’s gotten a lot more complicated since then. One of the things is bid management always used to be on a keyword by keyword basis. They dealt really well with high volume keywords, but not so well with real long tail keywords. Now the tools have gotten a lot more sophisticated in handling problems like that.
Kevin Lee: Sure. Well, that’s just one of many challenges that the designers and engineers working on bid management technology have worked on over the last several years to try to make the technology better at making the right decision, right? The decisions bid management technology has to make are actually fairly simple; leave it alone, bid it up, bid it down.
Of course when bidding up or bidding down; it needs to decide by how much. How it gets to that decision is what ends up being critical, right? Is it looking just at its own data; is it looking at its own data on a weighted moving average? Is it using historical data to try to be more predictive; there is a lot of stuff that will really hurt your brain if you try to get into it.
That’s why I mentioned that you have to try to figure out whether the type of business you are in is a fit for the kinds of algorithms that these particular engineers have written into the platform. Has the vendor had to deal with a particular problem for another client?
That’s where the innovations come from, challenges that a particular marketer of sufficient size faces that a tool vendor or technology vendor is actually willing to develop the energy into creating a customized module for them.
Eric Enge: Right. Any other types of tools that you would recommend that people look at?
Kevin Lee: Data in general can be really useful. So, tools that allow you to get a better sense of the competitive landscape and those vary from tools like Adgooroo, SpyFu, Hitwise, and comScore. Those are all tools that give you a better sense of the competitive landscape. And then, there are tools that the engines are actually starting to roll out and improve, so Microsoft for example has a huge initiative with their “KSP services platform.
Then, they’ve got a spreadsheet application that you can use called AdSage. You are basically trying to make an educated guess before you spend the money to find out that something didn’t work, so the more relevant data you can examine, the better. You can guess that something will work, and be a little bit closer to accurate.
It saves you that testing cost, because everything that you roll out in paid search or any media really for that matter has a testing cost associated with it. It’s the winners; the blockbusters that pay for the losers just like the movie industry. But, you want to try to increase your chances that you are going to have more blockbusters.
Eric Enge: Thanks for taking the time to speak with me today.
Kevin Lee: Thanks so much for having me; I certainly appreciate it.