Steve O’Brien
The following is the transcript of an interview of Steve O’Brien, VP Internet Solutions Marketing for Unica, and co-writes a blog with Rand Schulman of Unica. Steve is responsible for driving Unica’s success in the online marketing space. With over 20 years of management experience in the technology industry, Steve’s career includes director and VP-level roles in marketing and business development at Sun Microsystems, Red Brick Systems, and Personic Software. In 1999 he helped form the original management team as VP Sales and Marketing at Fireclick, a Web analytics provider acquired by Digital River, Inc. in 2004. He was most recently VP Marketing and Business Development at KickApps, a Web 2.0 social networking infrastructure provider.
Steve holds an MBA from the Johnson Graduate School of Management at Cornell University and a B.S. in Operations Research & Industrial Engineering from Cornell University.
Interview Transcript
Eric Enge: Can you provide us with a self-introduction?
Steve O’Brien: I am relatively new to Unica. I just joined about two months ago; been in the industry for about twenty years. I actually started out of Business School with Digital Equipment here in Marlborough Mass. After that, I spent most of my career in Silicon Valley. I moved to Mountain View in 1989 to work for Sun Microsystems. And, I worked for Sun as a product manager and product marketing manager for six years. I worked in Commercial Market Development Group, working with relational database vendors, specifically a company named Oracle.
I left Sun in 1995 to join Red Brick Systems. Red Brick was a data warehouse company. We built a relational database for Decision support applications. Working with Oracle, and working with the Red Brick was my introduction into the whole business intelligence or knowledge management space. Red Brick was a good run, and we went public, and the company got acquired after that. And then, I left to take a VP of marketing role at a company called Personic Software. Personic was a traditional enterprise software company selling into human resources applications, a compliment to PeopleSoft.
It was only after that in 1999 that I hooked up with the founders of Fireclick, which was a Web analytics company that we sold off in 2004 to Digital River. I have been in the Web analytics space, specifically since ’99, but in the Business Intelligence space and Decision Support for probably ten years, or twelve years total.
Eric Enge: Can you talk just a little bit about what Fireclick was about, and what made it unique?
Steve O’Brien: Sure. Fireclick didn’t start out in life as a Web analytics company; we started as an Internet Infrastructure Company doing content acceleration. We had this unique algorithm for caching content in the browser cache, so that no matter what site you visited if it was Fireclick enabled, it appeared to be very, very fast. It appeared like you are on a broadband connection which was impressive in 1999 when broadband connections were rare. Of course, as web servers got better and faster, and web developers got more talented and broadband became more ubiquitous that the need for that type of technology went away.
But, we found a different use for our IP in that we were tracking every click from every visitor to try to put these models together of usage patterns. And, some of our customers at that time asked us if they could tap into the knowledge that we had built about the visitor’s usage patterns. We then came out with Path Analysis. We actually pioneered a few of the technologies that have been widely adopted and copied by others. Path Analysis was one that we came up with on our own; the Overlay User Interface, where you open a browser, and you see either a heat map or bubbles that show where people click on your website were both introduced by Fireclick.
More recently we introduced the thing called the Fireclick Index, which was an average across all of the Fireclick customers, which is dominated by internet retailers of things like what’s the average conversion rate for someone in the consumer product space; or what’s the average page view for a media site? So, the Fireclick Index is still available today at fireclick.com, and it just puts web analytics data into context so that if you know that your conversion rate is 1.8%, you can see how that compares to your peers in the industry.
We sold the company to Digital River Inc. in 2004, and unfortunately, they haven’t invested a lot in continuing to build the Fireclick business, so it’s sort of been left behind.
Eric Enge: Why did you decide to join Unica?
Steve O’Brien: I think the web analytics space is fascinating and fun; for me, it’s the perfect blend of my business intelligence and enterprise software background. I also like to be in a space where there is lot’s of hype, but working for a company that’s really focused on solving the business problems and exploiting the opportunities that are available. So, it’s just a really good blend of enterprise software; solving a real business problem, but still being involved in the Internet, working with exciting companies that are doing exciting things. One of the reasons I left Fireclick was that I thought that Web analytics was becoming increasingly commoditized, all of the products started to look the same, with all the vendors copying each other’s features and functions. Prospects will tell you that all the products look the same.
I was attracted to Unica because I think they have a good vision of where space needs to go. And, I think that they value the Web analytics portion of Internet Marketing, they know it’s a crucial piece, and that the web is a very, very important channel. But, it’s just one channel in a much larger marketing problem that companies are trying to solve. And, I think Unica has a good vision for how this space is going to evolve and shakeout, and I think they are pretty well positioned to take advantage of that.
Eric Enge: Great. Let’s explore that a bit. Can you give us a working definition of cross channel integration?
Steve O’Brien: Cross channel integration is exactly what it sounds like; it’s the ability for companies to provide the best possible customer experience for all of their customers, regardless of how they chose to interact with the company. Whether I call the support center, or order from the website, or order from a catalogue, or walk into a store; my experience as a customer and with the company should be fulfilling and seamless, and satisfactory.
Eric Enge: How does that play out, when you start talking about doing analytics and trying to model customer behavior across channels?
Steve O’Brien: It plays out because you are gathering lots of information about your customers every single day. Web behavioral data is well known, and most web behavioral data is anonymous. And so it’s like doing market research with a 100% sample size. Because, you are sampling every single person that comes to your website, so it would be like doing an opinion survey of every single person that walks into your store. And so, some of the data that you gather on the web obviously is anonymous, and some of it is not, because people do register.
People buy things and enter their name and address, and credit card number; they sign up for your newsletters; they join your social network. There is a lot of data available on a website through analytics that is not anonymous. But, it’s very difficult for companies today, or at least companies just starting to tackle the problem of integrating this cross channel data.
So, if I call the support line versus whether I visit the self help support website, or whether I order something over a catalogue, or order it on the website, or walk into the store in Downtown Boston; the company should know who I am. They should know their history of servicing me as a customer; they should have a record of, if not my preferences, at least a record of my past purchase behavior. And so, it’s integrating all those things across channels that provide a better experience for the customer.
Eric Enge: Why is that so important?
Steve O’Brien: Mostly it’s important because of its effect on the top line and the bottom line. Obviously, it affects the bottom line, because it saves money. Just think about a cross channel marketing program, if you are sending expensive direct mail to the same people who have already responded to your direct email, and those are the same people who were already loyal purchasers in your store, you are just wasting money on marketing programs that are targeting the wrong message to those people.
It makes absolutely no sense to have overlapping, redundant or conflicting marketing programs. And so, those very simple examples that I mentioned are good examples of why you just want to have an integrated cross channel marketing effort. It affects the top line, maybe in a more meaningful way, because all the customer communications combined from your company to the customer; determine that customers total experience. And, the total customer experience determines how much people buy; how frequently they return; how loyal they are; and whether they recommend your business to their friends. And, if you think about the biggest trend or the biggest buzz word today in marketing is word of mouth marketing; WOM. And, the best way to generate good word of mouth marketing is just to provide a superior customer experience. So, an integrated cross channel presentation to your customer base is the best way to generate, how is it that customer experience?
Eric Enge: Right. It was a couple of scenarios that come to my mind; one of them is that you have a person who is ready to buy, and they have already decided they are going to pay the full price. And then, in the mail, they get a coupon.
Steve O’Brien: Exactly.
Eric Enge: Right. Or similarly; and just it is well, frustrating for the customer because that situation I just told you isn’t frustrating for the customer. That the one that is frustrating for the customer is; they already bought the product, and they paid less price.
Steve O’Brien: Right.
Eric Enge: And, now the email comes chugging along and saying, hey 20% off; that’s right; I bought it yesterday.
Steve O’Brien: That’s right.
Eric Enge: And so, these are scenarios where you are causing consternation, or turn with your customers because you are not managing where they are. I don’t know how you do anything about knowing that person is ready to buy, because they haven’t told you that, you wouldn’t know that. You send him a coupon and well, that’s life. But, if they have already bought a product, then you really shouldn’t be sending him an email with the coupon; here it happens all the time.
Steve O’Brien: It happens all the time; and in fact I mean you may decide that if you could eliminate that occurrence; if you could eliminate the offering the discount to someone who just bought the product yesterday; you just created an unhappy customer out of someone who had just bought something yesterday.
Eric Enge: That’s right.
Steve O’Brien: Instead, if you just said; we can identify who those people are, and we want to offer them something special, and increase their satisfaction. So, we offer them 10% of their next purchase; or we will offer them a free upgrade. But, you can still offer those people something; you can still have an effective marketing program, but the message is completely different.
Eric Enge: That’s right. And, now you are building your business rather than thrashing your business.
Steve O’Brien: Exactly.
Eric Enge: So, what are the key technical challenges for an analytics company in pulling all this together?
Steve O’Brien: It’s not rock it times really; Web data is really well understood by all the analytics companies basically. And, most of all the analytics companies provide fairly good ways for users to consume, or analyze, or distribute data either in reports or in some web-based interface. And, companies that our customers have been integrating; desperate applications for a long time, but it’s always a pain point because the vendors never make it easy, and this could be the topic for another three-hour interview. But, we all know that the frustrations that customers express, when they try to get two products from two different
20.01 vendors to work together; there just always is a hurdle there. And, so whether it’s through open systems or common file formats or standard schemas, or published APIs; it is completely technically feasible today to integrate almost any two enterprise applications. If there is a good enough business reason to do it; and cross channel marketing is a good enough business reason. The only technical challenge is really self-inflicted by companies and vendors who take a fairly proprietary approach to their data or their services. In Web analytics specifically, because all the leading Web analytics packages are; most of them are delivered via software as a service or host of applications. It does add a layer of complexity to data integration. How do I get the data out of the host to the application? What data, am I allowed to add to the host of the application? How does that flow own the data; what about data quality, data integrity? There is just a whole host of issues around integrating; hosted application data with non-hosted or non-managed data. So, that’s where some of the complexity arises specifically to the Web analytics vendors.
Eric Enge: Right. And, don’t you also have situations where even after you’ve completed the integration of the volume of available data is very different. Like you might have very simple data coming in from some offline source; compared to the huge rooms of data that comes in from Web analytics. Doesn’t that complicate things or maybe it’s not so much an issue?
Steve O’Brien: It doesn’t have to complicate things for me; you are right in that web analytics tends to produce a lot of data very quickly; even sites that generate modest amounts of traffic can generate tons and tons of web data. But, integrating that, it’s all in the schema and who designs the integration; it doesn’t have to be a big challenge.
Eric Enge: Right. So, let’s talk about it from the website owner’s perspective, or maybe that’s the way you answered it before, or maybe not; let’s think about that. Yes, there is two perspectives in my mind; there is the challenge that the; let’s say Unica has, as an analytics company in pursuing cross channel data integration and then there is the challenge that the website owner has.
Steve O’Brien: Right.
Eric Enge: And, I think you really addressed the website owner perspective, so I am not mistaken. Can you also address what’s involved in having an analytics company have the technology where result to work with all these different solutions; different types of data?
Steve O’Brien: Sure. And, they are completely related, right. Because, the analytics companies have all taken different approaches to these integrations from do nothing, to offer some finite list of prepackaged integrations that are guaranteed to work to publishing common APIs, and inviting partners to program to those APIs to. Some vendors even have completely opened to backhand systems that can integrate with anything, and that’s how it’s back to Unica’s approach. If website owners need to work within the confines of the solutions provided by the analytics vendor, then their first challenge is to choose a vendor that integrates cleanly with the other applications he has. So, either choose a vendor that has an out of the box integration with your most important other application whether its email, or campaign or search bid management or whatever it is. Choose the vendor that has the best integration with that product or choose the vendor that integrates pretty openly and easily with all other applications based on some published standard, and not a proprietary vendor-specific standard, but some open standard. And, we’d obviously prefer that everyone did the later, and if everyone picked vendors that integrated easily with everybody else; would be pretty well positioned.
Eric Enge: Right. That would make life easier for the customers as well.
Steve O’Brien: It would.
Eric Enge: Alright. So, what are the various types of cross channel data; you’ve partially listed this before, but if you could provide a focused with that would be great?
Steve O’Brien: Sure, Yes. So, I mean most of our customers; member of Unica Service Marketers; so most of our customers do direct marketing. And so, most if not all of them use direct mail or email; so we see companies integrate email and direct mail whole of the time. And of course, email is almost always
25.01 integrated with web behavior or web data. But, we have also seen companies integrate call center data with online support data; so both Web data and call center data typically that’s to try to drive more customers to the website because it’s a lower cost of service. We have also seen companies integrate Web data with campaign data, so that people who exhibit certain online behavior receive specific offers via email or mail or from the telemarketing group. So, I mean its call center; its database; its prospecting customer database data and its campaign data.
Eric Enge: Right; alright. And, do you have some implementation examples that you can talk about, that would illustrate some of these things?
Steve O’Brien: Sure. I’ll give you a couple; I am not sure if am allowed to use their name.
Eric Enge: That’s alright. We don’t really need the names in the high-level example I think; they are great.
Steve O’Brien: So, they are very, very large company; I spoke with the other data used; web analytics data, combined with their internal customer database and their direct mail list. And, they used the combined and cleaned up list to drive attendance at their annual user conference.
They use email marketing techniques to drive what they called early bird registrations. And then, they integrated the online registration data with the web behavior information, to better target people who they deemed as likely attendees. Then, they sent mailed invitation, hard copy expensive invitations; only to people who hadn’t yet signed up, and also set the profile of likely attendees, so they saved money. They also optimized the landing pages to offer better offers to the folks who responded to the page search campaigns. And, then they modified landing pages to maximize their conversions; so they were able to exceed their registration goal for the program and they spent less money doing it.
Eric Enge: Right. So, analytics software plays a role in keeping all this together?
Steve O’Brien: Yes. They used NetInsight; the unique product as sort of the main data repository for all of this information. So, that’s how they were able to blend the campaign data with their customer database, with their registration data of people who had actually responded and converted. And, also all the traditional web analytics practices of optimizing landing pages and measuring the cost and the effectiveness of their page search campaigns.
Eric Enge: Right. So, and basically, they gave people doing this, a single place to go where they could get all the data starting with the analysis of who would be the best match for the user conference. And, deciding how to break that all out; they could do that just by getting into the analytics interface, and it’s just doing some basic analytics type analysis.
Steve O’Brien: Exactly right. And, they could all do all that to do a single interface. They didn’t have to go to difference; let’s go look in the registration database. So, let’s go look at our installed customer based database; let’s go and look at who visited the website database; it was all integrated into a single database.
Eric Enge: Oh Yes. You’d have a very serious headache, going through all that stuff. Yes, cool. So, you said you had another example; you might be able to get.
Steve O’Brien: I can think of one more; there is a company called Synergy, and they are among the many companies that figured out that their call center service agents cost them on average; I think it was four dollars and fifty cents per customer call. And Yes, the customers that visited the self-service portion on their website cost less than a quarter; maybe they said something between four cents and twenty-four cents. I couldn’t put a number on it, but it was basically attempts or less of the cost, so that is actually five percent.
And so, they looked at their call center data integrated with their online support sender data, and they were able to target segments of customers who use their call center but tended to not use the self-service portion of the website. And then, they did a targeted email campaign to those segments of customers who overused the call center and didn’t use the self-service website, and they explained how easy and fast it was. And, how much better the service was, and they got an immediate response over their website
Eric Enge: No call-wait time.
Steve O’Brien: Exactly. All the great advantages of using the web and they realized that for every successful conversion; every person that went to the self-service website instead of calling them; there was a ninety percent part of the lie.
Eric Enge: And, the obvious part is you are driving someone from a four fifty or whatever it was to less than a quarter; that’s really straightforward. But, the other thing that happened there is, they selectively mailed only those people who were problematic.
Steve O’Brien: Exactly.
Eric Enge: So, they won’t send it; people who are regularly using the web interface already, didn’t get this email.
Steve O’Brien: Right. So, our previous example of; hey I already used this self-service; still I am saying what are you telling me about it for. And, that’s why integrating the two data sources was important.
Eric Enge: So, with that let’s go to our wrap up question, which I always like to wrap up an interview with. Major challenges facing the analytics market in the next year.
Steve O’Brien: Wow! That’s a big question.
Eric Enge: Yes. Well, we are going to finish big.
Steve O’Brien: That is big; I think there is a lot actually. I’ll just rattle them off so we can dive into whichever one you think is worthwhile. But, I definitely see and hear from customers and prospects about this increasing; what’s called commoditization of the web analytics space. And, what I mean by that is, the lower end products are getting better and better and the higher end products aren’t getting that much better over time. And, that it’s very difficult to discern the differences between something that you can get for free, and something that costs hundreds of thousands of dollars. And so, this commoditization is really driving prices that vendors can charge for their products and services down.
And, while vendors are looking for ways to offer more values for customers, and to also to differentiate themselves from their competitors. And, you see all the vendors taking slightly different approaches and different routes. And so, I think it will be really interesting over the next year to see how all that evolves. And, I think everybody is placing that’s in a different area as every vendor is trying to expand their suite of products and services in different directions. So, everyone is trying to figure out what’s the direction that makes the most sense; what’s the one that’s going to have the most customer attraction, the most resonance out there in the market.
Eric Enge: Right. How do we not get swallowed by Google analytics?
Steve O’Brien: Exactly, exactly. And, it’s not a trivial problem. I mean it’s not a no-brainer; there is a lot of really smart people that think about it every single day, and nobody has got the definitive answer.
Eric Enge: Right; Yes. Now, that’s an interesting one, others?
Steve O’Brien: Others I think, every vendor talks about Web 2.0, and almost every web analytics vendor that talks about Web 2.0 wants to talk about event tags and Rich Internet applications. I think the really more interesting aspects of Web 2.0 are the social aspects, and being able to track them; and measure, and monitor, and track the social aspects of the internet is going to become more and more important. So, are people talking about my products and services in their blogs; are they linking to me from their social networks; are they; have me in their RSS reader, it’s just…?
Eric Enge: And now, you are measuring things that aren’t on your website?
Steve O’Brien: Exactly. How do you do that? I mean how do you measure the word of mouth marketing on the internet? I mean it should be easy; there are all these little startups popping up; Buzz Monitor and Buzz Metrics, and all these companies that report on the overall customer impression of you on the internet. And, I think their room for web analytics vendors to help people measure their presence across the internet, not just the traffic on their website.
Eric Enge: Yes. I think that’s a very interesting topic area by itself which, and you are diving into forms; you are diving into blogs; RSS like you said. I mean not only how many people are subscribing to the feed, but also…
Steve O’Brien: Reading the articles in the feeds that they are subscribing to.
Eric Enge: Yes. And, how many of them are rendering data from the feed on their website.
Steve O’Brien: Great; which is then being read by…
Eric Enge: Others.
Steve O’Brien: Exactly.
Eric Enge: Yes. And these are multidimensional; you have to crawl the whole web to measure that.
Steve O’Brien: Theoretically; unless you come up with a smarter way to do it. There’s got to be a better way to do it, right. Every company is not going to be able to crawl the entire web to look for references to their name or their brand; so there has to be better ways to do it. Someone will figure it out; we will figure it out.
Eric Enge: Yes indeed. So, also as you said, a lot of people like to talk about rich internet applications and the issues with that. And, I’ll just mention the one thing that I heard, which was the most interesting way of looking at that, which is that the notion that the measurement part is straightforward. It’s deciding what you want to measure and therefore present to people, what’s the metric people are looking for here, and that is the harder part. Does that make sense?
Steve O’Brien: Yes. I think Eric Peterson writes pretty intelligently about event metrics, and I think that is the equivalent of KPIs for events. And, I don’t believe that every rich internet application event is equally important. But, I think the challenge of figuring out which events to measure, and which events to monitor, and which ones matter; it’s about the same order of magnitude as figuring out what are the right KPIs. And, I think once one smart person does it, and writes a book about it; everybody else can tap into that.
The only different challenge there is obviously, that these applications will continue to evolve, and we don’t get them today, and we think we understand Rich Internet applications, and we look at AJAX, the dynamic HTML, FlexiFlash, and we think oh Yes, we understand that. Two or three years now that will be completely different, and they will be using different technologies, and they will be doing more amazing different things that we can’t even f. So, it will keep changing and evolving in a much more dynamic way then KPIs did.
Eric Enge: Right.
Steve O’Brien: I mean I don’t know what year I brought the first KPI book, but people are still reading it today; still referring to it.
Eric Enge: Yes. Web Analytics Demystified.
Steve O’Brien: Exactly. And, I’d say the biggest, just back to your original question; it’s a really boring answer, but I think that number one major challenge that I hear about from customers and prospects is web analytics are great. I have got all the data in the world, but I have no idea what to do with it; I have no idea how to interpret it, or how to analyze it, or what action I should take based on it. And, I see just a huge need in this market for expertise; people who know; who understand Web data and understand how to get to the root of what to do with the data, and how to use that data to improve the business.
There is not a customer I talked to that doesn’t struggle with that.
Eric Enge: Right. Well, there aren’t enough good analysts out there for one thing.
Steve O’Brien: Exactly.
Eric Enge: The other problem is that not everybody does a good job of planning for the need to have a good analyst.
Steve O’Brien: Well, that’s true! Many companies think, oh, I’ll just go get analytics in a McDonalds, Pizza Hut, type way, and that is a problem. The problem wasn’t that you didn’t have the data; the problem was you didn’t know what to do, and just because you bought an analytics tool, you still don’t know what to do, unless you’ve got somebody smart there to help you to figure it out.
Eric Enge: That’s absolutely right. Great, thanks.
Steve O’Brien: Great. Thank you!
