Eric Enge interviews Live Search's Grad Conn
Published: December 10, 2007
As senior director, product marketing, Grad Conn is responsible for both the enterprise and consumer sides of the Health Solutions Group (HSG) at Microsoft Corp., including market mapping and segmentation, brand and communication strategy, and all marketing communication and sales enablement implementation.
In the enterprise and healthcare provider space, the HSG offers IT solutions that link disparate systems to maximize their value and enable providers to make informed decisions to improve patient outcomes and lower costs. In the consumer space, the HSG is developing new tools and services to empower consumers to manage and share information with trusted parties for improved health. These combined efforts are part of the HSG's vision to help people around the world live longer lives by enabling them to make the best health decisions with the aid of IT.
Conn brings to the healthcare industry 20 years of experience as both an entrepreneur and executive in the software and marketing fields. He has a unique background that combines nine years in the consumer packaged goods industry at Procter & Gamble and 12 years running self- and venture-funded software companies in the peer-to-peer, social networking and bookmarking spaces. These companies include OpenCola Inc., which was named by Fortune magazine as one of the 25 coolest technology companies of 2001, and SiteShuffle Inc., which was declared The Next Big Web Thing at ETech 2006.
In his personal life, Conn has been heavily involved in theater for many years as an investor and producer. He is a board member of the National Music Theater Network Inc., the parent organization of the New York Musical Theatre Festival; the 5th Avenue Theatre (Seattle); and the Broadway Educational Alliance. In addition, Conn is one of the principal architects of the Nederlander Alliance LLC's Audience Rewards program, the first loyalty rewards program for Broadway.
Conn holds a bachelor of commerce (Honors) from Queen's University, Kingston, Ontario. He is married with two daughters.
Eric Enge: Can you provide an overview of what your thoughts are on the health related search market?
Grad Conn: eMarketer reports that online advertising market in the pharmaceutical and healthcare business is about 800 million dollars right now. It's expected to be 2.2 billion dollars by 2011, and so that's something like an annual compound growth rate around 17%. That's a total number for all online advertising: display, search, lead, and all the other aspects of online advertising.
About 40% is usually deemed to be search, towards search driven revenue, and 20% would be display. Then about 18% will be classified, and then the numbers get to single digits after that in terms of rich video, and lead and a whole bunch of other stuff.
It's a pretty interesting market, where search is the dominant form of advertising, and it's growing quickly. Today not that many people do that much with their health online, versus baking or travel. As that starts to change and more people start to do more health related activities online, I think everyone's expecting we are going to see a lot more product and service activity online as well.
Eric Enge: During the Searchification event you held recently where you announced your latest update, you said that 80% of all search users have done health related searches online.
Grad Conn: That's a number from Jupiter - of online users, 80% have done a health related search sometime in the last two or three years. That said, the percentage of queries from health searches is not the highest category. The breadth is one of the highest, because everyone's got some kind of relationship to a health issue at some point. So, the number of people doing searches on health is very high, so the reach is high, but the frequency is a little bit lower. In fact there was some new numbers I saw on eMarketer just yesterday.
They show that 89% of people are researching specific medical conditions and illnesses; so it's a pretty high incidence. What we find interesting about it, is that it's something which is an activity that is broad, and when people are doing it, it's very important.
Eric Enge: It's usually a big priority at that moment.
Grad Conn: Well, it's not necessarily a priority, but it's important. There are really three conditions people are in when they are searching. One is that they have a symptom, and they want to know more about that symptom. That usually is a situation with a lot of emotion.
Another condition is the diagnosis, so I have been told I have something. You see a lot of emotion, and a real desire to learn more and understand more, and to try to figure out how to get better. The third one is actually where I have a resolution. I resolve to lose weight, I resolve to eat less; I resolve to stop eating candy. People go in there and they actually are trying to change their lives in some way. That's a little bit less urgent, but there is a different kind of passion around that.
Eric Enge: What about the rate of growth of health search, you alluded a bit to it in the initial data you gave me from eMarketer. Is the rate where health as a percentage of all searches is going up?
Grad Conn: Well, that's a good question. I am not sure that as a percentage of all searches, there is evidence that it's going up right now. So, the idea being that search is in itself increasing very, very rapidly, health search is going to increase overall, as part of it.
Eric Enge: What about the user's reaction when they do a health search and they see ten blue links for results. How has that worked for them?
Grad Conn: The key challenge in health search is that when people start it, they often don't know what they don't know. And so, they'll start with a relatively generic term, but there is a lot of complexity in health search that's quite different than if I am looking for a digital camera, or booking a flight; or I want to get Britney Spears latest pictures. The level of complexity in it, the type of information I need to discover is quite challenging. I may have some kind of diabetes, but it's not going to be the same as everyone else's.
I am going to have even different issues to think about, and different things that I am concerned about trying to solve online. Maybe I really want to understand what kind of exercise program I should be doing if I've got diabetes, and that's what I want to use the web for. It's a very different kind of set of results than if I am trying to understand whether I might have diabetes, because I am thirsty all the time.
Those are both very varied searches, and if you just type in the word diabetes and get back ten links, it doesn't cover it. People are quite frustrated with it, because then they have to turn themselves into researchers, and really start figuring out how to construct queries, and the ands, and the ors, and there is this weird process of trying to figure out how to trick the engines into giving you what you want. What a lot of people are doing in health is creating ways of helping people get an overview of all the different concepts that fall underneath that general topic, to help them refine their search in a useful way, and get to the information they want faster.
Eric Enge: Right. You must have done a lot of testing just on the base experience, not only on your modified versions.
Grad Conn: Well absolutely, tons of testing as you can imagine. What we see is that people get frustrated quickly in the general search engine experience, and they tend to abandon it. They'll go sometimes to message boards and other places and gather people's help. But, they don't really know what to do, because many of the results that get returned are articles that are reasonably complex and sometimes difficult to understand.
What we ended up doing is we've created what we call a dashboard. As I've said everyone is trying to be helpful in health search; you see almost everyone doing something. What most of them have done is they created topic headings like physicians, conditions, diet, and other very general topic headings. We did something a little bit different. We expanded those individual headings as well, and the way we expanded them is that we looked everyday at which journals have been published by the National Library of Medicine. We look at peer-reviewed journals, and create an index of those journals, and that index creates associations between terms. So, if a new article comes out today that shows there is a link between diabetes and computer use; I hope not, but, if that happened, we would know that tonight, and then tomorrow if I type the word diabetes in, computer use could be one of the topics that would show up in the dashboard. What we are doing is actually getting a bit more granular with the drill downs in that dashboard to allow people to understand more, faster.
Eric Enge: Right. One of the big things you did is you made this notion of refinements dynamic. Many months ago Google Topics was announced; actually I think it's over a year ago now. But anyway, that was a human edited version, which of course inherently is not dynamic in structure. You've taken this extra step and made the whole processdynamic.
Grad Conn: There is no right or wrong way in this, right? Everyone is trying different ways of making this work, but we think that in health there is a bit of a difference in that, the opinion of experts is quite important. By using peer-reviewed journals, we are effectively creating a virtual expert index. And, without actually going to experts and asking them to index the entire body of medical knowledge, we are doing that virtually by going through their peer-reviewed journals.
I think that's a pretty powerful way of thinking about it, because overtime concepts change as well. And so, experts will have a very strong point of view, and then ten years or fifteen years later they may have a very different one. So, grapes are good for you, grapes are bad for you, cantaloupe is good for you, cantaloupe is bad for you, etc. We see that happen all the time and we know how to track that, and see how that change occurs.
Eric Enge: Right, at Searchification you talked about how you did user testing. Part of that involved moving in with people and bringing cameras? You called it Experience Maps, I believe. Can you talk about how that went, why you did it; why it was important, and the quality of data you get out of that?
Grad Conn: Sure. There are two basic kinds of research, qualitative and quantitative. Most qualitative research is done with focus groups. The great thing about it is you quickly hear peoples' points of view, they are reasonably inexpensive and quick to pull together, and you can get a sense of what people are thinking.
The challenge with focus groups is that they are subject to group thinking, and they may go in different directions depending on who is the most talkative in the group. And, they don't always drill down to really important emotional issues, because it's a group of strangers meeting each other for the first time.
Hence, people will often use the qualitative as a way of creating the framework for a quantitative study which is one on one. It goes much deeper, but again it's quantitative, so people are completing a survey. The strength is you get a much larger base, and you've got stuff you can segment. You can do stuff with the numbers, and you can drill down; but again it really lacks real emotion.
Experience Maps are a technique that Procter & Gamble's been developing over the last ten years. I worked at P&G for nine years, so I am familiar with the technique and with the company. When I got here it became clear that health is a very emotional subject, arguably the most emotional subject. It's tightly connected with our fears around mortality, and tightly connected with our hopes and dreams for our families. So, I wanted to make sure that we were correctly understanding and reflecting that emotion.
It's very easy in a company like a training company in the IT industry to get locked into the technology, and to get lost in the mechanics of making that stuff work. We want to make sure that we are actually reflecting what real people want in their lives. So, we did do focus groups, we did do a quite a bit of quantitative work, and then we took what we learnt from that. We identified groups of people in the Midwest, what we call "average Americans", and got them to try to open up to us. The way we did that is, we had a camera crew and a set of researchers call on these families. It's like a focus group, but a bit more one on one, because it's just our group and the family.
Overtime, the familiarity of constantly visiting, and constantly filming, and constantly being part of their lives, creates a really easy flow back and forth. What you can see coming out of these people is that these women that we call the family health manager, are really focused on how do they protect their family? They see themselves in this role of protector, and it is a very important role for them to fulfill. That was a bit of a new insight for us. We tended to look at things in a very results oriented way. They wanted a system in a way of being able to protect their family, and provide a family environment that enhances their safety and security on a long term basis.
Eric Enge: One of the things you implemented was the inline article results and the three-column layout. This is a very interesting notion, this notion of integrating the article results in here directly, and I presume your usability testing showed that a large percentage of these kinds of queries are people who really are looking for authoritative sources to answer their questions?
Grad Conn: Yes, but I wouldn't say that that's the full motivation for what we did. The real motivation goes back to a comment I made a bit earlier. When people start doing a health search, they don't know what they don't know. And so, the dashboard is helpful, but the dashboard is obviously single words or two-word pairs, it's not super descriptive.
What we see people do is they would like the dashboard, but sometimes if it was a newer topic or something they weren't familiar with, they needed a little bit of a runway to understand the topic a bit more before they could fully use the dashboard in a way they are comfortable. Those articles are meant to provide an overview of that search term with that as your topic, and then what people do is that they see this interplay between the dashboard and the articles as they learn very quickly. Before they click in any web results, they are actually able to educate themselves, and refine their queries significantly.
We've got different types of sources in there. What we find is people really, really like getting stuff from the Mayo Clinic. So, we've got content that, we've contracted with the Mayo Clinic, and they are providing articles in there. Some people really want to see what the official government position is, so we have got the NIH stuff in there. And, some people really love Wikipedia, now that's somewhat controversial at times, and not everyone agrees to that etc., but I know I have come from a community of people who love Wikipedia, and would be horrified not to see it there, so Wikipedia is in there too.
Over time we are going to continue adding new sources in there that allow people to look at the point of view that's important to them. I know that thing is not in there right now, but really should be is, what does the rest of the world think? So, the World Health Organization, NHS in the UK, and other health systems around the world can be incorporated. Maybe I also want to get the Chinese medicine perspective on this issue. The thing about medicine is there is not one single point of view that's correct.
There are many points of view, all of which consider themselves correct, and I think it's important to allow people to see them, and choose amongst those. We are not here to tell people what the right medical point of view is, we are here to deliver the information and they can choose what they want to see more of. And, there will eventually be a feature, where people will be able to indicate that they are really not interested in anything from a particular source.
Eric Enge: They can begin to tune their experience.
Grad Conn: As long as they are a signed-in user.
Eric Enge: It's interesting if you go back up to the refinements, you have a category called alternative medicine. You talked about providing different points of view, and it's got three things listed in the results, such as green tea, flavonoids, and more. It's intriguing that flavonoids made it to the top of the list. I presume that which ones you show are selected by some sort of popularity, visibility on the web metric.
Grad Conn: Two comments, You can just click on show more refinements than it will actually blow it out a bit more. But, these are based on peer-reviewed journals. So, that's actually their ranking here; how often under the heading of alternative medicine, the term cancer and flavonoids appear in relationship to each other.
Eric Enge: This notion of inline articles; is that something you think you can apply to other verticals?
Grad Conn: I can't speak with any authority on other verticals.
Eric Enge: Just to speculate for my own sake, it would seem to me that if there were other areas where you could identify similarly authoritative sources, that you could apply this to other verticals. You'd have to find ones where it really applied, but I think it could be quite intriguing. You also talked a bit about the privacy concerns when searching on health topics. Can you talk about that a little bit?
Grad Conn: Health is a pretty sensitive area for most people and depending on what you are searching on, you may not want everyone to know that you are looking at that. Privacy and security for their health information is a key concern for people. Most consumers we talk to, and this came through loud and clear in all the types of research we did, view their health data as the most important data that they have; more important than their financial data, more important than anything else. Partly because, once it gets out, it's out there.
Whereas, if I lose my credit card, it's a big problem, but I can fix that. I can get a new credit card. I am not liable, life goes on. If my employer finds that that I have AIDS, life might change for me in an important and significant way; and that's not something I want to live with. That's where health search becomes quite different. We found that people were asking us a search in health that was more secure, so we've encrypted it. They wanted to be anonymous, they wanted us not to be tracking them, and it was a pretty broad understanding of the amount of tracking that's going on in the search business today.
That's where we made a decision to dump the logs after ninety days, and not keep the cookies. Some privacy advocates have been super pleased that we did it, but also surprised from a business standpoint. But, we think it is the right business decision, because creating an environment where people feel comfortable and safe is the environment people will want to do health searches in. There is more to come there, so there is a lot of other things you can do to increase the privacy and security, and other things to make people feel more comfortable with health searching, but we think we've got some of the big ones in place already.
Eric Enge: Right. Certainly there has been a lot of discussion in the media of late about the privacy concerns and the lengths of search engine cookies. It's intriguing, that discussion on it own. I also noticed that overall in the right, in the sponsored results above the normal MSN ad center listings, there is a box with books from Amazon. It looks like a special advertising relationship that they struck with you.
Grad Conn: We felt that Amazon had a really nice structure. One is, they've got a lot of books that are relevant to health topics, and they have a lot of products, mainly from our partners that are relevant as well. So, we've got those in as effectively an ad union on the right rail.
Eric Enge: Right. I guess there are certainly times where people want to get a book, and then the answer that they are really looking for is going to take that level of research. Last but not least, how's the reaction then? Do you see signs of increased engagement, or increased satisfaction with the results they are seeing compared to what you saw before?
Grad Conn: We are seeing double digit increases in search result page views, which is a great indicator, because they are doing more searches in that spot, so they are seeing more results and doing more queries. We are also seeing triple digit indices in the growth of the searches, but it's on a small base right now, so I am not going to get too excited. We are on track to hit our traffic numbers that we were hoping to get in the first few months. So right now all systems are green, and we are continuing to move ahead. We've got lots and lots of changes in refinements and improvements to make.
So, in a very typical Microsoft way, we get stuff out there, test and improve, test and improve; continue getting consumer feedback. We've got lots of people giving us comments, the folks we did all the Experience Map work with, have been actually playing with it, and we've been getting lots of good frank feedback from them on what they like and don't like. They've had a bit of a hand in building it, so it's quite exciting to watch them get engaged that way.
Eric Enge: Right. So, you are doing testing now on an ongoing basis, now that you have a platform open to the public.
Grad Conn: Yes, there are little changes slipping in all the time. There are new features in here that we have implemented since we announced this.
Eric Enge: Do you see a lot of people clicking on the refinement links?
Grad Conn: That's what drives the higher search result page views. People use those to ladder down to a more specific search. What people will do in a general search engine is they will type in new terms in the search box, and they'll keep adding terms in the search box.
With refinements they don't have to use the search box after entering the initial term. They use the dashboard to refine it, and they are effectively building a longer query, but they are doing a few clicks as opposed to typing and trying to think of the words on their own.
Eric Enge: Right, in this environment that is a better experience.
Grad Conn: Well, yes. I can guarantee you that if I was doing a search for cancer, there is no way in a million years that I would type in cancer and 5-FU. But when I see that I can click on it to see what it is. I can get that refinement, and I can understand what it is. Now, I am interested in that and I can click that, and my process of learning is just rapidly sped up, because it's giving me the taxonomy for what I should be thinking about when I think about cancer.
I might be thinking when I go to cancer about the right diet. So, I might type in cancer and I see under nutrition that dietary fiber is the thing that's most often related to the word cancer in the peer-reviewed journals. Well, I should probably read that first, and I may not have thought of that on my own. I probably wouldn't have even expressed it that way as a user, I probably would have said maybe fiber or maybe roughage or something else, but then this is the way it gets me to the answer a lot faster.
Tomatoes are number 4 in the refinements. I would never have expected tomatoes to be number 4 in nutrition around cancer. It's fascinating to watch people, because everyone's got fairly unique conditions. And, these family health managers, what makes them special is they are not just managing their own health, they are managing not just the health of many people, but the health of many people in widely different age ranges.
They are dealing with infant care, and child care, and teenager care; and all the issues there. They are dealing with older parents who've got a whole host of health issues as well. These family health managers are actually managing a very complex set of health conditions that even today's health professionals aren't really exposed to anymore, because they tend to specialize. This kind of stuff is great for them, because they can really quickly get an understanding of what the categories are.
Eric Enge: Right. Ultimately in these situations you are talking about, when they go to see the doctor, they know what questions to ask. Certainly the great majority of doctors do a great job, but even the good ones are extremely busy people and making sure that the most important concern you have is addressed is a very important thing.
Grad Conn: Our technology is from a company we acquired in February called MedStory. It was built by Dr.Alan Rappaport. He is a medical doctor, and he built this because of that. He had patients were going online and doing searches, that wasn't an issue. But, what they were coming back with was not helping them. He decided to give these people a tool that allows them to be better educated. He also decided to create a tool that allows doctors and allows physicians to drill down really well, because just a fact of life the world is, we've got a rapidly increasing body of knowledge about our bodies, and it is impossible as a practitioner to stay up to speed.
The way doctors are compensating for that is they are specializing, and increasing their narrowness, their focus, so they can at least try to be up to speed in a specialty that they exist in. But even then, it's always useful to have someone come to you with something that you may not have read, not because you're not a good doctor, but because there are only so many hours in a day, and you can't see everything that's published in the world.
Eric Enge: Right. Are there other aspects of the new health search that I haven't touched on here that you'd like to comment on?
Grad Conn: I think we've covered most of it. The only thing we didn't talk about was the scrapbook. On the top right hand corner there is a scrapbook, and the scrapbook allows you to save a whole bunch of different things. You can save ads that you see, individual books from Amazon; any of the links that come up in the web links. The articles that come up in the Article Viewer, and something we just added recently is the actually query string can be saved now to your health scrapbook.
What we find is that health is different because of the confidentiality, and health is also different because often health conditions that you are researching are recurring conditions. Occasionally, you just have something you are dealing with that day, but often if you are going to spend the time online, it's because you've got something you are going to be dealing with on a long term basis. So, people like to be able to save the things that they've found, to be able to go back to them, and to be able to refer to them; to share them with their doctor, that kind of thing. We launched this on October the 4th. What the scrapbook does is effectively store the health information, which in this case will be search results into a private health storage platform called healthful. So, that's why healthful search and healthful, same names all tightly tied together.
Eric Enge: That seems useful as well. You want something private here, and you're dealing with a variety of different kinds of objects at the same time.
Grad Conn: Yeah. The thing about healthful is that it's totally controlled by the user. The user controls the records, what goes in, what goes out; who it gets shared with. Maybe I am doing research on something I don't want other people in my family or my workplace to know about, I can store that in a way that I feel very comfortable that I am storing it safely and confidentially in a way that no one else can get access to it.
Eric Enge: Right, well great. Grad, thanks for taking the time to speak with me today.
Grad Conn: Great, well thank you. I enjoyed it, and hopefully that was helpful.
Have comments or want to discuss? You can comment on the Grad Conn interview here.
Other Recent Interviews
- Microsoft's Grad Conn - Dec. 10, 2007
- Seth Godin - November 26, 2007
- Blogline's Eric Engleman - Nov. 19, 2007
- Microsoft's Ramez Naam - Nov. 12, 2007
- Microsoft's Garry Wiseman - Nov. 5, 2007
- Twitter's Jack Dorsey - Oct. 15, 2007
- Google's Matt Cutts - Oct. 8, 2007
About the Author
Eric Enge is the Founder and President of Stone Temple Consulting (STC). STC offers Internet marketing optimization services, including SEO, Social Media and PPC optimization, and its web site can be found at: http://www.stonetemple.com.