Dennis Mortensen
The following is the transcript of an interview of Dennis Mortensen, COO of IndexTools. Dennis Mortensen joined IndexTools Inc. in 2004 and he is responsible for identifying market opportunities, developing and executing direct client and partner strategies for the company including professional services and technical support. He is Principal of Elisabeth Victoria Corp., a company specialized in funding the most early-stage, high-growth companies in the online advertising industry as well.
Dennis Mortensen has extensive experience in web analytics and online advertising. He began focusing on web analytics in 1996 when he founded Canvas Interactive, a web analytics and Internet consulting company – which was later acquired by a pan European public company. Prior to joining IndexTools he has operated as an executive at several high growth Internet companies.
Dennis Mortensen holds a bachelor’s degree in Computer Science from Skive Business College and is an accredited Web Analytics associate teacher at British Columbia Institute of Technology. He is a recipient of the best young manager award in Denmark in 2001.
Interview Transcript
Eric Enge: I understand that you have been involved in analytics since 1996. That would make you a crusty-old veteran.
Dennis Mortensen: Exactly, I was there before, under, and after. I was even one of those who were lucky enough to sell their company in April of 2000.
Eric Enge: That couldn’t have been too bad.
Dennis Mortensen: That was fantastic timing, NASDAQ above 5000. So, yes, I think I have seen it all. I focus on the commercial side right now but spent the beginning of my career in development.
Eric Enge: Okay, great. Can you expand upon this, and provide an overview of your background, in a little more detail?
Dennis Mortensen: We started the company, on June 1st, ’96. And back then, this wasn’t IndexTools; this was a different company with a different focus. We didn’t even call it Web Analytics back then, we actually called it Statistics. And if you remember back then, Statistics meant something as basic as a counter on your page, which said that you had something like ninety-eight visitors to your website.
But it also meant log file analysis, and this is actually where we started; offering log file analysis, as the solution. We sold the consulting together with the software, and we used WebTrends; actually even before WebTrends had a version number. So, we used WebTrends back in ’96, and since it was software, we kind of had what I would call a fake on-demand model. That meant that we downloaded the log files of some three hundred clients down to our servers, we analyzed the log files overnight, and then we presented the results of those log file analyses reports either in weekly, biweekly, or monthly meetings. We would take all the analysts and actually go out to the company and present it. And, we did that for four and a half years up until 2000. When we sold the company; we were some fifty consultants.
Eric Enge: Impressive.
Dennis Mortensen: That was from the consulting point of view. We have actually seen that coming more and more back right now. But what we sell today is a product. We are doing a subscription model where people pay for use of the tool. 95% of the company revenue is based on revenue from the product. A smaller portion is based on consulting provided by our professional services department. So that means, we as a company, we don’t do SEO, or SEM, or anything else to do with online marketing we provides a Web Analytics tool, and we provide professional services upon that too. So I have spent five years on the consulting side of Analytics and now five years on the tool’s side.
Eric Enge: When you think about the IndexTools product, what do you think are your key differentiators?
Dennis Mortensen: I should, of course, give you the grand one hour, sales and marketing picture about how great I think we are. But, being brutally honest here, I think we are moving towards a market, where the tool itself is becoming a commodity. I don’t really see much difference in between Omniture, for example, and IndexTools, besides that Omniture is in green and IndexTools is in blue. I still think there is a difference in the market, especially since Google Analytics advanced a little more than a year ago, meaning that everything below Google Analytics kind of disappeared.
So you don’t have a hundred and fifty tools anymore, it’s more like twenty tools. That means anything above Google Analytics is probably split into two categories: the enterprise vendors and the SME vendors. And, IndexTools is definitely part of those six or seven enterprise vendors. Looking at those seven, I think the differences are on the backend, I don’t mean to refer to whether or not a vendor uses a pre-developed database like Omniture; I think they started out on MySQL or something like that. At Indextools, we actually developed this from the bottom up; we actually have our own database. But that’s not really so fantastic, but the fantastic part is that we can scale it, and we can scale it extremely cheap.
The second difference is ease of use. We started out in the SME market and talked to marketers who weren’t fully educated in online marketing as well as you see larger enterprises being. So, we have been forced to create a tool, which you can log into and start using instantly because we have been serving clients at $10 a month. And at $10 a month, you can’t really add that many hours of professional services. So the tool itself had to be easy to use because we couldn’t afford to have too much support on it.
So we have an easy to use the tool. And then, the third part is price. As you know, everything in the industry right now is priced on volume not really features, some of the competitors have extra charges for segmentation or data warehousing, and so forth. But, it all comes down to one thing, volume. How much volume do you have, and that volume is usually measured in page views. That means the solution costs less if you have one million page views a month; and costs more, if you have one billion page views a month. And we are the most aggressively priced enterprise tool out there.
We are probably one fourth of the price of Omniture, one-third of the price of Coremetrix, and one half of the price of WebTrends. So, none of the other enterprise vendors can beat us on price; even though we have the same functionality. So, these are the three points that we go to market with: a back end which is extremely scalable, ease of use, and price.
Eric Enge: So you see IndexTools competing directly with Omniture, Core Matrix, and WebTrends, as opposed to ClickTracks, or somebody like that?
Dennis Mortensen: If I were to write up a comparison map without having a hundred sales meetings, I would probably have included NetTracker, which was just acquired by Unica; I would probably include Instadia, or Site Intelligence, or ClickTracks, or some of the other contenders.
But in the last ten customer meetings, we met the top four every time. It may be that Omniture gets invited to 5 times as many meetings as we do, but it’s always, Omniture, Coremetrix, WebTrends, and WebSideStory that we end up competing with. We rarely meet ClickTracks, Site Intelligence, Instadia, NetTracker, or any of the others.
Eric Enge: Fair enough, and you talked a little bit about the market segmentation, so in your view, it really is a two-part segmentation, just small-medium business products versus enterprise type products?
Dennis Mortensen: You are actually not that far off there. We have kept the attitude of being able, not from the technology point of view, but from an organizational point of view, to serve the SME market. That means you can start at $50, and then move up. We can start you from a hundred thousand page views, but you can move up, and we have clients above a billion page views a month. We can actually serve you all the way through.
If you take my three points on where we are different; we are definitely different on price, so there we can grab very high volume clients, those who probably are fewer marketers, but have a high volume site. These are typically ideal customers for me. I tend to go after marketers who don’t necessarily have a fulltime web analyst on board, because if they don’t have that, they actually have a marketing assistant or themselves to work the tool. The ideal customer is anybody who wants a detailed analysis; that means you are in a specific industry and want something very unique for you. You are actually able to pull that information out.
Eric Enge: Are there some of your larger customers whose name you can share?
Dennis Mortensen: Sure: PriceRunner, the shopping comparison site, ToysRUs, Vodafone, and Tesco, which is the same as your Wal-Mart.
The EMC storage company as well and the mobile content operator; it’s called Jamba in Europe, and I think you call it Jamster in the US.
Eric Enge: I think that the common problem we see in analytics as a whole, is figuring out who wants what information, and making sure they get it. For example, a large company like ToysRUs may have many different people who want different things. Is that a common thing that you have seen?
Dennis Mortensen: Yes, I agree with this. We have clients, who have no web analyst on board at all, who have only a halftime marketer on board, and we have clients who have a staff of six people on web analytics only. But I think that a lot of people make the mistake thinking that the tool is going to solve the problems for them, or the tool itself is going to be the savior of their web solution; it is not.
The tool itself doesn’t really do anything; it’s the analyst. You can set the tool up to do certain reports, to investigate something, or to alert you on specific metrics; but the tool itself doesn’t really do anything, it’s a reporting tool. You need some kind of analyst, or you need some kind of objective that you measure on and take action on, before there will be any return on investments on a web analytics tool. This is actually part of our sales pitch. If you are going out and spending $100,000 on an Omniture solution; that might be your budget. You can actually have an IndexTools solution for $30,000, and hire a web analyst for the remaining $70,000, and you will probably have ten times the effect on your website results by doing that.
Eric Enge: This is, of course, one of the big things that people talk about now, business intelligence, and how you use your tools to extract the intelligence in a way that jacks up your business. Are there any aspects of the IndexTools products that are a pretty nifty business intelligence features that you would like to highlight for us?
Dennis Mortensen: I think, unless you have really a high volume site, you would not see the difference between the different tools. If you have a high volume site, you are not really used to having real-time analysis. If you have a billion page views ongoing month-over-month, having the ability to do real-time analysis, it’s not something you would expect. We have clients who move from, not necessarily one of our competitors, but from their own data warehouse, or from their own business intelligence tools, where they do one report a month.
That’s the speed at which they can aggregate data. And, we have the ability to do reports and do analysis across metrics in real time on billions of page views; and that is actually pretty unique, and this is something that really impresses people. The latency of our system is three seconds; that means, from somebody actually visits your site till they are included into your segments, or filters, or whatever you have set up, they are there. You don’t have to reanalyze, you don’t have to do anything, three seconds later the data is there.
Eric Enge: It offers a nice piece of functionality when you are just trying to test your setup.
Dennis Mortensen: You and I kind of expect that in our low volume web sites, but enterprise clients certainly don’t expect that. They expect monthly or quarterly reports. And all of sudden, being able to have something in real time, not within the same day, but within seconds, is amazing to them.
Furthermore, being able to actually to use it as a business intelligence tool, such as implementing segments on multiple metrics without having them set up in beforehand. For example, if you use Google Analytics, which is a pretty neat tool, you have to set up a segment in advance, and you start to fill up that bucket with data.
In a tool like IndexTools you can set it up, and look at data in the retrospective, and see what actually happened last Christmas for these specific keywords, for this specific campaign, for this specific set of uses, and see who actually did buy something. And you can actually set that up and say well, I want to repeat this activity this Christmas. I think there is an immense amount of power in being able to do that. But it, of course, comes back to needing somebody to do this for you, you need the analyst.
You need the analyst to dig into the data. A lot of companies see e.g. affiliate marketing as a decent way of doing marketing, but they just don’t have the affiliate marketing skills. So, you will see a lot of these companies outsourcing the affiliate manager, for example to a network, or to an agency. And I think you will see some companies outsourcing their web analyst to either web analyst specialist companies, or agencies, and so forth.
Eric Enge: Business intelligence experts.
Dennis Mortensen: Yes. I tend to go with the word web analyst for now; I think business intelligence is probably too big or too strong a word for what we are looking at right now.
Eric Enge: What role can an analytics package play in terms of addressing click fraud, and what are the key things that need to have to be effective in doing that?
Dennis Mortensen: First a personal comment, and this is not based on how I feel, or who I work for. This is based on aggregated data across some four thousand clients that we have. If you read an article in Newsweek about click fraud being in between twenty percent and thirty percent, that’s just plain ridiculous, and it is wrong, and I think to click fraud, in general, is hyped. I do agree there is click fraud, and I can make click fraud in a second, I can go to Google right now, and start doing it. However, the picture painted is not that bad in reality.
Coming back to your question; first of all, the analytics tool needs to have search engine modules, which can collect information from the engines. You need to choose a vendor who has verified API agreements with Google, Yahoo, Microsoft, Miva, and so forth, which we have.
That means, we can get all the data from the engines and power that with the analytics data, and so the first thing you need to be sure of is that the analytics tool actually can collect all this data. And when you are sure of that, you need to make sure that you have set up the specific actions that define success on your site, so you can determine whether we are talking about click fraud, because I think we can agree that there is no such thing as click fraud if you have been successful in a conversion.
So you have to look at anything that’s not successful, and I think that the analytics tools, not just IndexTools, but any of the other analytics tools out there are doing a better job of addressing click fraud than the click fraud tools themselves. First of all, because we are extremely customizable, so click fraud can mean anything to any company; so with us, you can actually determine when you think fraud has happened.
That can be x number of visits without conversion, or x number of visits from specific IP’s, or domains, or referrals, or that could be x number of visits who had a specific path, and it’s easy to set that up. So, you yourself can build an indication of what click fraud is. Furthermore, in a tool like ours, you can set it up to send out alerts, saying that anything above this threshold indicates some kind of abnormal activity.
Not many of our direct clients utilize the tool to pick up on click fraud, but most of our agency partners actually have set up customer reports to at least assure that the abnormal traffic is kept under a certain percentage. And so, if they actually find something which they think is click fraud, they actually use the tool itself to provide specific reports to send to either Google, or Yahoo, or any of the other engines.
Eric Enge: What percentage of the page clicks would you guess, or what the range you would guess that click fraud represent?
Dennis Mortensen: In between two percent and four percent.
Eric Enge: So it’s a fairly small number, I mean it is not that it shouldn’t be addressed, but is not what some make it to be.
Dennis Mortensen: Correct. At two percent, on an $8,000 a month PPC budget on Google, how much time can you actually spend on finding potential click fraud, and requesting that refunded from Google? Not really that much.
Eric Enge: That’s right. So there is a related issue, which I referred to as invalid clicks; so let me just give you an example of what I am talking about. An invalid click might occur when you have said that you only want traffic from the United State, but someone comes in through a corporate network, and they are actually working overseas somewhere, show up as a local click, and then click on something that’s not really any potential for them to buy. That I don’t think as click fraud, I think of that as being an invalid click.
Dennis Mortensen: I think there are actually more invalid clicks going on than real click fraud.
What happens, either instantly, or at night, or the following week; for example, is that Google filters their reports for potential click fraud and potential invalid clicks. So even though your analytics reports might show the clicks you may not have paid for them.
And that’s why you need a vendor who has API access to the search engines, so you can actually match those things up and consolidate the data.
Eric Enge: Can you talk about the advantages and disadvantages of using a custom Javascript based approach to categorizing pages as opposed to using a UI based approach to try and categorize pages?
Dennis Mortensen: I actually think this is one of those small things that almost seems too obvious to be important, but this is one of the really important things in your web analytics deployment. First of all, many sites change their pages to optimize results on a frequent basis. If you change the URLs or Titles of your pages and you don’t use built-in Javascript labels, you are prone to error.
If you change your pricing page title four times over a year, the tool itself will think this is four pages. But if you label your page in the Javascript and don’t change those labels when you update the title of the page, that solves this problem. I think it’s important to take your more important pages and have a naming system in the Javascript for those pages on a document name basis like homepage, or category page, or whatever.
Setting up document groups gives you better usability because since we offer a real-time system, you can still group different pages together in the segmentation system, or in the filtering system later on. But if you set up the document groups from the very beginning, if you want to look at a segment of those who visited your electronics category, it’s easy to do instantly. You could still pinpoint those twenty-eight pages and do a segment based on those twenty-eight pages later. That becomes difficult if you have a hundred thousand pages.
So definitely, I think grouping pages together will be better for filtering, and it will be better for segmentation. It will also be better for path analysis because sometimes you don’t really care where the people move from the Acer TravelMate’s 8 product page to be Acer TravelMate 8.1 product page, what you really want to know is when they move from Acer to Sony, or from Sony to Dell. And that might be a reasonable path that you can look into, where the other specific paths are just too unique for you to really take action on. This is another case where labeling document groups in Javascript is valuable.
Eric Enge: Right. Let’s talk a bit about first-party cookies. I have seen many different kinds of implementations. For example, creating custom C-name records in the DNS, listing all the domains and sub-domains that might be part of the site we are looking at, or simply doing nothing at all. Can you comment on this?
Dennis Mortensen: We have seen a deletion rate of 10 to 15 percent on third-party cookies among our clients which is certainly large enough for any client to move up and start using first-party cookies. We set multiple cookies where possible – this meaning that beyond the first party cookie we also try to set a third-party cookie. Our clients have to install the tracking script code on their servers as part of an IndexTools deployment and it’s from this script the first party cookie is set. This makes it a real first-party cookie, set from your server, and your domain.
This also means that if you use any third level domain names like http://www.yourdomain.com v.s. http://register.yourdomain.com, it all goes under the same first party cookie. But should you have another second level domain as in http://www.yourdomain2.com we try, based on the 3rd party cookie, to track the user across different 2nd level domains and if the user allows 3rd party cookies we register this cross-domain path and update the 1st party cookie at the same time?
Eric Enge: Can you talk a bit about metrics-based approaches versus report based approaches?
Dennis Mortensen: The big thing here is the attitude in the way you collect and present data. The approach that we have taken is more in line with what you see in more traditional business intelligence tools like Cognos. We offer a metric based approach. We don’t try to give you a canned report on every single metric that we provide, such as what browser people use. What we try to do is collect metrics, that we don’t have pre-canned reports for.
This means that for deployments and use of the system you need to know what you are looking for.
I personally think this is a much better approach than report based systems. WebTrends for example, provides something like three hundred reports or more, hence you spend your time looking for that magic report, for that to solve your problems, and you start in the upper left corner, and you work your way down, and you end up doing what I call report surfing. You spend one and a half hours surfing reports, they are all nice, they are all fantastic, they are beautiful, but it doesn’t really drive action.
I would rather have started with a business question saying something along the lines of, I need to know whether the $10,000 I invested into this specific campaign drove any activity within this merchandise category, and what the profit was – okay; that’s a business question. We provide you with the tools to rapidly design a custom report. This is a faster and more efficient process for most businesses.
Eric Enge: What do you see as the key challenges facing the analytics industry in 2007 and beyond?
Dennis Mortensen: I actually think it will be less to do with technology and more to with methodology. One big question is how a web 2.0 application including Flash application or any other rich internet application will be tracked. We need to agree on what it is and what we are supposed to track here. It’s easy to track events – note that events seem to be the new word for what happens within the same page on web 2.0 applications.
So, how many events do I want to track, one, two, ten, twenty, every click or movement? Currently, there’s no real agreement in the industry. It is going to be challenging to agree on what it is that we are supposed to measure because we can measure everything now, but what is it; that I, as an online retailer with a web 2.0 property is supposed to measure?
I think that’s going to be exciting. And then, I think we will see the integration of a lot more external data sources. People have pretty much come to understand the web data that we collect today. Now, they want to see more data either within the same tool or have the ability to export this data out to another tool. The reason for the analytics package typically being the data holder is that, we are the ones with the most data; because we collect tons of data compared to some of the other channels. So, they tend to import data from other channels into our data store.
Eric Enge: Like call center data, for example.
Dennis Mortensen: Exactly, or any other specific data that they collect, that affects their business, It’s not just about web data; it’s going to be about other data as well as you indicate.
Eric Enge: And what about rich media type solutions, like mobile phones, or AJAX, or about you know, different kinds of things that aren’t currently so friendly to Analytics collection?
Dennis Mortensen: I don’t think it comes down to the technology part, because they are easy to track. But it’s very difficult from a methodology point of view, to agree on what are we supposed to track here, and I think what we are going to see is a lot of people actually just trying this out. Mostly for optimization purposes, such as AB testing, multivariate testing, or something along those lines. But I don’t think, unless an industry organization such as the Web Analytics Association, says “an event is this”, that you will see any agreement about how to proceed, and all the vendors will do their own thing.
Eric Enge: What do you think the key questions businesses should be asking when they are deciding upon analytics solutions at this point?
Dennis Mortensen: I think they have to ask themselves the first question, which is whether they actually have some well-defined business objectives for their web properties. It can be something as basic as we want to make money, fair enough, then I can help you define key performance indicators (KPIs) around making money. You can, for example, have a website that generates leads for two markets, each market expects different things.
Good, when you have that, you can go out and ask your analytics vendor, how much of this can they can track, and how they can present the results! Do this instead of saying “show me what kind of features you have”. Because we all have fantastic presentations and fantastic features, which can dazzle everybody. But the out of the box reports and presentations don’t really mean much unless all you want to do is track very simple key performance indicators. If that’s all you need, Google Analytics can track that – so why pay $10,000 to get something you can get for free?
You need to define your objectives internally before you start the process. If you do this you will get so much more out of the tool you choose, and you will get a much easier deployment by being a little bit prepared by knowing what it is that you want to achieve.
Eric Enge: This makes a lot of sense because there are so many things you can do with these tools today, that it can be bewildering.
Dennis Mortensen: Exactly, you end up, and I am sorry to repeat myself, report surfing, you do basic tracking, and you look at the some of the reports, and they are fantastic, and they all look nice. But, you don’t really drive action. What I try to pitch to our clients is, please don’t focus on any of the canned reports that are there – they are nice, you can use them as templates, but don’t use any of the canned reports. Focus on using the custom reporting capability to get reports based on metrics that you use to drive results for your business.
If you set up a custom report, it’s because you have a business question that you want an answer to. If you look at a canned report, it’s because you don’t really have anything else to do.
Eric Enge: Right, excellent, it’s been very interesting.
Dennis Mortensen: Yes, thank you!
