Brad Goldberg Interview
Published: August 4, 2008
As General Manager of the Search Business Group, Brad Goldberg is responsible for product management, product planning, and marketing for Live Search. This includes current and future versions, and spans all search verticals like Local and Image Search, as well as customer segments including consumer and enterprise. Prior to this role, Brad was the General Manager of the Windows Client Product Management Group where he led product management, launch and planning for the Windows Vista as well as Internet Explorer, Windows Media Player, and Windows Update Services.
Brad joined Microsoft in 1997 as a Product Manager in the Server and Tools Business Group and managed the launch of Visual C++ 6.0. Following the launch, he led product management for Visual C++, .NET and C#. He then moved to the SQL Server Group where he led business strategy and product marketing for SQL Server. After the launch of SQL Server 2000, he became the Director of Enterprise Strategy, where he managed efforts around sever pricing and licensing, enterprise sales models, solutions marketing, and business planning. Next he became the General Manager of the Server Platform and Enterprise Product Management Group where he managed strategy development, marketing, and sales integration for the entire portfolio of Microsoftís application platform, including SQL Server, Visual Studio and BizTalk Server.
Before joining Microsoft, Brad worked in management consulting as an Associate at A.T. Kearney in their Chicago and Tokyo offices focusing on strategy development, mergers and acquisitions, and manufacturing operations across the automotive, utility, electronics, and transportation industries. He also worked in Corporate Development for the Coleman Company and as Assistant to the President for Nippon Tectron, a small Japanese custom electronics manufacturer.
Brad has an MBA from Harvard Business School and a B.A. in Economics from Amherst College. Brad also attended the Inter-University Center for Japanese Language studies to complete his post-graduate work in Japanese language.
When he is not working, Brad enjoys cooking, photography, and outdoor activities.
Eric Enge: I want to start off asking you what the grand strategy for Live Search. The last ComScore report was about 9.2%, which is actually up a bit for Live.com. Is there anything you could share with us about the strategy for growing your market share?
Brad Goldberg: Sure. First there is a lot of volatility month-to-month in share, and while we are always happy when we see our share increasing, we donít try and manage share numbers month-to-month. Instead, we try and focus a lot more on the core long-term things that we need to invest in, which covers three things. We need to deliver great results, which includes all of the work that we do around relevance and continuing to grow the quality of the search results that we provide. So that first piece is really just built around delivering great results.
Second is the idea that we need to simplify the key tasks that people do. We think there are four key tasks that people use search for today. They use it for entertainment, videos, commerce that they are researching and for purchasing goods offline or online. Search is a critical part of that process. People use it for reference to find a specific piece of information, whether it is sports score or a map, and they are just using the search box instead of the address bar as the vehicle for doing that.
We think that search is going to become more and more task centric, and that we need to really focus on how we can increase the role that search plays in simplifying these key tasks
The third strategy is based upon changing the model around search today. We need to try and create new value for consumers and new value for advertisers so that they have a different proposition when they look at using us or advertising with us.
Eric Enge: Right. So, changing the model around search is a mouthful by itself, isnít it?
Brad Goldberg: Yes, but we actually think we need to do all three of these things because, as weíve talked a bit before, it starts at the core with relevance. You need to be a great horizontal search engine to really start differentiating in key areas even within the four tasks that we are outlining. So it is about improving relevance and picking out key areas where we think the user experience can really evolve.
Eric Enge: Right. So, I was at the Searchification Event back in September, and one of the things I remember was the discussion of the four verticals. You broke out to sub-sector things like health and medical for example, where you created a new kind of experience.
In the commerce area there was the whole engine for integrating reviews into a master-guide of attributes for consumer electronics products, and giving them attribute-by-attribute rating scores and the like. You were really trying some pretty aggressive new takes at redefining the experience on a task basis even then. So I take it that the plan then is to keep doing those kinds of things and going further and further.
Brad Goldberg: Thatís right. Thatís what we talked about at Searchification, the start of us taking a more task-centric approach around how we looked at activities, and we have been on a really consistent strategy. We first started talking about it last September, but it has really been what we have been executing on from an engineering perspective for the last year and a half.
We are starting again with relevance, having that in place and then starting to pick these key tasks where we think we can differentiate around the user model. In September we first talked about the strategy to specifically amplify commerce and we introduced cashback as a core area when we think about the product experiences as well as around the business model.
We needed to highlight both those things; you need to do a great job with the user experience and you really need to think about commercial relevance in a somewhat different way than you think about , relevance for a breath of queries. There are different tools and a different experience that we need to bring to bear. And then, we have also tried to supplement that with a unique value proposition for both the advertisers and users around cashback.
Eric Enge: Right. So how much of a difference does it take to unseat an incumbent search engine? Just being 10% better is not enough, is it?
Brad Goldberg: No, thatís right. Within search it is actually very easy to drive trial. So when we look at different things from a marketing perspective, it is relatively easy to say we encourage users whether it is off of MSN or things that we are doing off our network. The point that you are hitting, about switching taking longer is completely right.
It is not just a function of how much of a difference you need, but it is also sustained over a period of time. If you think of the search behavior that users have, their queries will frequently be spread out over different verticals and they might be searching for certain terms at one point in time and then different terms the following week.
Most users are using multiple search engines today when you look at the user data, so thatís good. But we still need to build out that level of differentiation, so we are focusing more on commerce within a specific area. Then we can start to ask if we can get to a place of greater differentiation in that area that allows us to really carve out a place where people are starting to use us first for a set of transactions that they are doing online.
And also for more and more of their research that they are looking at when they are looking to purchase goods either offline or online? And then you hit a point where people decide to make your engine the primary engine. And so, it is not a linear magic bar where you say, here is the point that you absolutely hit. It needs to be a set of things that we do.
Right now we are trying to get pretty scrappy, pretty creative, and do some things to encourage trial. Over time we are confident that if we keep at it, users will want to switch over to us for a set of areas where we think weíre better and differentiated.
Eric Enge: Right. After all, the user doesnít sit there with Google, Yahoo, and Live search all lined up, and then compare the value of the ten blue links. That is because they are on a task. They are on a mission of some sort. As you said, it is not just about impressing them once, youíve got to get them over and over again, and this is why the differentiated experience makes a difference.
Brad Goldberg: Yes. I think about two different modes that people are in when they search. There are some users who are very active with how they search, they have a very specific intent and they have a preferred search engine that they use for all searches that they do online, and thatís a very conscious choice that they are making. There is also a wide, vast set of queries that are much more passive, where the user might have a toolbar installed on their machine, or it might be their default homepage.
In those cases, users are not consciously choosing a search engine. They are treating search much more as a convenience and not an experience. So for us to think about differentiation in growing our share, we really need to have strategies that pursue each of these.
Cashback is a strategy that is really starting to build more active preference for our search engine. Distribution deals like the one we did with HP are more focused on how we build out our passive share and get broader distribution in trying for entry points for our search in a broader number of places than we have today.
Eric Enge: Can you expand upon cashback a little bit and how thatís fitting into the commerce strategy?
Brad Goldberg: Sure. So it starts with the same view that queries are going to shift to be more and more task centric. We looked at commerce as the place where we wanted to amplify this focus first. It is about 30% of all queries depending on how exactly you account for different types of research queries, but it is a much larger percentage of the overall revenue. It allows us to do some more innovative things between the user experience as well as the advertiser experience. That allowed us to create this model around paid engagements that we rolled out with cashback, which creates new value for users and then new value for advertisers.
Over time, the success of our strategy and commercial search needs to be as much around the productís experience as it does around the business model itself. We need to go in and create new user experiences similar to what weíd started at Searchification last year. Not only things like the opinion index, but also how we think about evolving the overall shopping experience that users today conduct via search.
We need to think about what we are doing at the backend with commercial relevance so we are doing a great job not only with offers that we might ingest, but also with offers we crawl across the broad web. Together we create incentives for consumers and advertisers to think about doing things in a different way than they do today.
Eric Enge: Right. So, we talked a bit earlier about basic relevance as a component of this. And at Searchification, there was a lot of discussion about the relevance testing that you did prior to that release. Can talk a little about how you think about you measure relevance?
Brad Goldberg: Sure. The way that we think about relevance is it is something that there is only going to be a couple of companies in the world that we think are going to be able to invest in over the long term and be able to build out the data center infrastructure and the investments required. .
We think it is an absolutely critical part of being competitive over the long term in search. Microsoft is one of a couple of companies across the world that is willing to take that much of a long term perspective to be able to do that. In terms of how we measure it, human judgment is a pretty standard approach that most search engines use today.
One of the unfortunate things about relevance is that there is not a more standards based approach to it. There is not a clear objective way, where you have a third party who looks at relevance across different search engines and talks about who has made gains in what ways. As we step back and think about where we are today, our belief is that for the majority of queries, users canít tell the difference across search engines.
This is what we see in our research using human judging, itís what we do in third party research that we conduct. After Searchification there were a number of sites and blogs that went out and did their own takes to try and look at the difference across search, and they generally found results that were consistent with that. And so, while there will be some queries where we are better, there are some where Google is better also.
For the majority of queries, we think today that business has matured to the point that users canít tell the difference across engines. And that was really where we were trying to get with Searchification, and why that was such a key milestone for us.
Eric Enge: Right. So another thing that I think was talked about is that something like 40% of user search queries go unsatisfied. Thatís a big number. What do you think is achievable here? I mean, how much better can we get?
Brad Goldberg: Well, I think we should look to get better every year. The specific number is the percentage of queries that go unanswered where the user re-queries, and that is something that today is one of the challenges. So many queries are new and unique over a period of time. There are queries that search engines might not have seen before. So before May, there werenít a lot of people who had been querying for Iowa flood, and then you have a period of time where that becomes a popular search term that may tell them what is happening in the news at that point in time.
Over the long term, it is really hard to imagine how much better things could get than they already are today. But certainly, I think there are several changes that we can make in terms of improvement on relevance. Think about the role that natural language has on semantic technology, and weíve announced an intent to acquire Powerset.
We havenít fully closed on the deal yet, but we are in the process with them. Thatís part of our long-term investment and part of our belief that there is really going to be about a number of different approaches and a long-term investment to really continue to improve relevance.
On relevance, it is not that you will be 5% better and people will notice. Youíll have to be significantly better, in much the same ways that Google was when they first gained a lot of usage, where users really noticed there was a significant difference in the relevance that they were seeing. So the bar of relevance is quite high
Eric Enge: Right, I understand. So, what about the plans for getting profitable? I saw the latest announcement, there was pretty good revenue growth, but the loss actually doubled roughly to 1.2 billion over the twelve months through June 30th. What do you think is the time horizon for that, and what are your plans in that regard?
Brad Goldberg: Yes, I think those numbers might be for online overall. We are in an invest mode on a number of different fronts. We are in investment mode from a capital perspective in terms of building out data centers and building out the scale and operational capabilities that we need.
We are in investment mode when I come back to the deals like the distribution deal with HP that I talked about. And then we are in investment mode when we think about what we are going to do to create differentiation for things like cash-back and how we are going to promote that.
So, at this point, when I think about where we are at search, we are in a place where weíre managing much more to look at what we are doing to grow our usage, and also grow our advertiser engagement and make sure our advertisers are having a great experience. If we are able to continue to grow our usage and we are able to grow our engagement with our advertisers, then the right things will happen to get us to a position of scale.
Search is ultimately a scale game and it is one of the reasons why we think there is only going to be a couple of players on a world-wide basis that are really going to be around over the long term in search. The stakes and cost are so high, if you think about it from just the pure economic perspective. We are at a point where we have higher costs than we have volume and there is even more that we need to, to build the infrastructure so we can deliver a great experience that will ultimately help get us to scale on the user side and on the volume side.
Eric Enge: Right. You are talking about billions and billions of dollars to build all this stuff out.
Brad Goldberg: Itís over a period of time that you are looking at building out the data centers and the infrastructure in distribution. And again, because search is a scale game, as you start to get more and more scale, your monetization improves, your results improve and you get into a cycle around things where the economics can get better pretty quickly.
Eric Enge: So, since you mentioned a couple of players a couple of times, are you hinting that Yahoo doesnít have the financial ability to sustain this battle?
Brad Goldberg: I wasnít trying to make this specific comment on any one engine other than fixed costs. Search really is a fixed cost game, and I think that the amount of fixed cost that is going to be required will continue to increase for the foreseeable future. Thatís one of the things that is kind of interesting when I look at all the startups in valleys that are experimenting with different angles around search. There are a lot of fantastic concepts that different companies are looking to prove.
But unless someone is going to come up with a way to capture and deliver results in a performance way on a global basis without needing to invest in capital, there is going to be a need for a major investment in capital that goes with it. I think search is going to look dramatically different than it does today in a couple of years.
So, as we enter the next phase of search, I think we will continue to see a convergence around fewer players that are really in the business of doing horizontal search, and that we will have a lot of players who are doing more vertical search in specific areas.
Ebay is search engine that people use for products. Expedia is a search engine that people use for travel. Facebook is a search engine that people use to find other people. And so, I think you will see search become more and more central. And then, at the same time, I think that is going to continue to converge around fewer players.
Eric Enge: Right. So with that verticalization, you can take it to the point where as a user you might have thirty different things you use for different kinds of searches or you might find that there are one or two players out there who offer you a lot of different experiences and capture a lot of what the vertical search players are trying to capture.
Brad Goldberg: Yes. That will be one of the most interesting things I think around how search plays out. Today, if you look at search behavior, search actually isnít good around verticals. In many respects it is not in the economic interest of a lot of vertical sites to expose more and more of their content to search engines because then they risk being aggregated in terms of traffic.
The question will be, what happens in terms of the relationship between these vertical experiences and horizontal experiences? So an example is, I am trying to sell a couch and I wanted to get a sense of similar couches and what they have been priced at. If you just use any search engine today and try and look on a specific model, I would be looking at Craigslist. I can get to a vertical site that might have a listing on it, but I donít get any breath of results until I actually land on that site.
So I actually think itís quite a tricky thing that will have implications as far as how the dynamics of search will play out. People may just use one engine for all of their needs, versus using different sites. We strongly believe that itís going to be more around multiple sites than it is going to be just one place for everything that you possibly need.
Eric Enge: Right, but you can see the argument for not wanting it to be thirty places?
Brad Goldberg: You know, the average user has a core set of sites they go to for the things that they need. The question will be for the things that they need to do, how much more does search become a central role in those types of experiences? You know Farecast is an interesting one when you look at travel, where it is really bringing a search centric approach to travel. Itís very different than anything that you are able to experience on either a search engine or on a travel specific site like Orbitz.
They have used search technology to be more predictive around whatís happening with specific fares.
The model is different in travel. As it plays out across different verticals it will relate back to the core horizontal search experience. How much of that thing gets subsumed within the core horizontal search experience versus how much of it is going to be around distinct entry points for the different tasks that people think about?
Eric Enge: Or, even if you are in the core search experience, but you enter a search term that relates to a vertical, whatís the experience in the results?
Brad Goldberg: Yes, and thatís gets in to not only relevance, but also how good a search engine can be at actually detecting the intent of the user. Today, tabs are a great example. Tabs are on every search engine, and they get basically 0% click through. Users have the ability on either a search centric homepage or on a portal where they can choose to narrow their search, but it is just not expressed in a way that users think about.
So it means that if you found a way that users could clearly articulate their intent, and if you did a great job on the relevance around it, then you can start to see how within a horizontal engine, you might be able to differentiate. But then again, today in the last ten years, search hasnít really changed at all visually in terms of the experience. You just have to ask if just the current forces in play will do that or is there going to have to be some change and disruption that will lead to a more fundamental change in the business model.
Eric Enge: As a last question, I did want to get your thoughts on the Content Ads program and what your plans are for that.
Brad Goldberg: Yes, weíve had this in beta last year and weíve actually released it. So, advertisers sign up with us and setup an AdCenter to run campaigns. They can choose to run it via search or via content ads.
Eric Enge: Yes. And, what sort of sites are you talking about? Does this include exposure on sites that are core MSN type properties as well?
Brad Goldberg: Yes.
Eric Enge: So, thatís a way for them if they have the best bids to get exposure on major MSN sites?
Brad Goldberg: Thatís right.
Eric Enge: Well, super. Thanks Brad!
Brad Goldberg: Hey, thanks very much for your time, Eric. Iím glad we were able to connect.
Have comments or want to discuss? You can comment on the Brad Goldberg interview here.
Other Recent Interviews
- Vitrium System's Peter Nieforth - July 28, 2008
- Aaron Wall - July 21, 2008
- Google's Carter Maslan - July 14, 2008
- Yahoo's Priyank Garg - July 7, 2008
- Izea's Ted Murphy - June 30, 2008
- Vanessa Fox - June 23, 2008
- Google's Matt Cutts - June 16, 2008
- Hitwise's Marc Johnson - June 2, 2008
- Bob Massa - May 26, 2008
- Yahoo's Andrew Tomkins - May 12, 2008
- Google's Dick Costolo - April 28, 2008
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.