00:01Alright, welcome, everybody.

00:02My name is James Killick. This is my colleague, Kyle Watson.

00:05We're going to have some Lightning Talks here.

00:06:…we're actually not going to show any slides. Yay-hey!

00:07The first one now is on optimizing site selection using Business Analyst.

00:12We've got one in 20 minutes, roughly, on using Business Analyst to optimize your marketing…

00:18…and then another 20-minute session right after that, Using Business Analyst to Optimize Your Territories and Districts.

00:26They're going to be Lightning Talks.

00:27We don't have a lot of time, so we're going to get started right away.

00:33This first one on optimizing site selection using Business Analyst…

00:40You don't have to be bored with slides.

00:42So we're going to get started right into some demos.

00:47We're going to divide the demos into two parts.

00:50One is optimizing site selection based on looking at an area that you've never been into before.

00:58This is very common with retailers, right?

01:01So if you've got an East Coast chain and you're moving to the West Coast, et cetera…

01:05…how do I look for a new location that's going to be good for me?

01:09And then Kyle's going to cut over and do some demos of how you can use Business Analyst to do site selection…

01:17…if you've already got a network of stores and how you want to fill in the gaps.

01:21Now, these two examples happen to be retail focused…

01:25…but don't assume that you can just use Business Analyst only for retail.

01:29It can be used for anything.

01:30It can be used for locating hospitals, government buildings…

01:34…you name it, the same data, the same techniques apply to the site selection process.

01:42OK? Any questions? Alright, we'll get started.

01:46So what I'm doing here is I've opened up Business Analyst Online.

01:50This is our web-based application.

01:52This is designed for anybody to be able to use.

01:55You don't have to be a GIS professional to use it.

01:59You can be any kind of professional at all.

02:04You don't, you know…

02:05…familiar with a consumer mapping product, you can probably use this.

02:08I've zoomed in to the San Diego area here, and I'll just zoom out a little bit more to get a bit of a broader view.

02:17There we go.

02:19And how many people here are from San Diego?

02:23One person. How [many] people know things about San Diego…

02:28…in terms of the kind of people that live there and, you know…

02:31…let's say you were going to open a store in San Diego. Where would you go?

02:34Let's say you were going to open a fast-food store. Where should I open a fast-food store here?

02:39It's a great question.

02:40Maybe the first thing you've got to do is learn something about the community.

02:46So in Business Analyst, you have a tremendous amount of data that you can search on.

02:51There's demographic, up-to-date demographic data, and I'll just search on some now, and we'll go to per capita income.

03:00And we'll look at 2010 income for the area, and we'll try and learn something about San Diego.

03:06So immediately, I get a nice color-coded map here.

03:09I can change the color scheme very quickly, and start to highlight where the high-income areas are on that map very, very quickly.

03:18I can dive into other data, too.

03:19I can dive into education, for example, and see where the people are that have a bachelor's degree who are over age 25.

03:29And there we've got another picture of San Diego.

03:32I can dive into the data and look at some age. Let's look at where all the infants happen to be concentrated.

03:43So we'll look at 2010 population, age 0 to 4 years, and here we're getting a slightly different picture.

03:50The high-income area was over here.

03:53The areas with high volumes of, or high densities of young infants are in different areas.

04:01So very quickly, you can start to build this picture up of an area…

04:04…and you can use the kind of picture to help you make decisions.

04:07It can be based on demographics, but it can also be based on consumer spending.

04:12So in this case, I've picked beer, and I'm going to look at imported beer, and in this case, I'm going to look at an index.

04:21So in a lot of cases, we have numbers, but we also have percentages of households that have a certain category…

04:31…like the percentage of households that have infants.

04:35In this case, I'm looking at an index where 100 is the national average, 150 would be one-and-a-half times the national average…

04:42…50 would be half the national average.

04:44So you can see these areas here in red, which is in the range 165 to 210.

04:50We just adjust those ranges slightly and I'll bring that one down to, say, 150.

04:56These areas in red now are the areas that are one-and-a-half times more likely…

05:01…to consume imported beer than the national average.

05:07I can also look at the data and start to zero in on lifestyle data.

05:17So maybe I'm looking at opening up a sports store, and I want to find out anyone who's interested in tennis.

05:24I could look at tennis, for example, here and see where there are high concentrations of people that play tennis.

05:31So that's very good. That's poking at the map.

05:34You learn a lot by poking at the map…

05:35…and you can explore an area and learn what kind of an area it is, but it's still poking at the map.

05:42So I'm going to move on to the second part of my demonstration now…

05:46…which is about how do you find an area that is right for you.

05:49If you're a business, like a retail store, or if you're trying to open up a facility, like a hospital…

05:57…you probably have multiple criteria that are important to you.

06:02So in the case of a quick-service restaurant, for example…

06:06…you're probably looking for certain characteristics or demographics…

06:09…and in fact, if you go to a quick-service restaurant and you say, "Hey, I'm interested in opening up a franchise"…

06:16…they'll say, "Well, in order for you to open up a franchise…

06:20…you need to have a certain population; you need to have a certain income…

06:23…you need to have a certain family size," et cetera, et cetera.

06:27How can you find those areas?

06:29Well, within Business Analyst, there's a tool to do that. It's called Smart Map Search.

06:35And I, actually, went and grabbed a Hardee's restaurant brochure from their franchisee brochures…

06:41…and I looked at the criteria that they had in the brochure…

06:46…and I actually entered it into the Smart Map Search tool that we have in Business Analyst Online.

06:52And this will come up in a second here, once I get the Internet to respond to me.

06:58They had certain income requirements, family size requirements.

07:02They had certain growth requirements and housing unit, occupation…

07:08…and I added one more variable that they didn't ask for…

07:11…which is the number of people that ate at fast-food restaurants…

07:15…specifically Carl's Jr. restaurants, and looked at the index of the people who did that.

07:21So their criteria was 30,000 for per capita income; family size of 3.1; median age of between 18 and 49 and a half…

07:31…and the unoccupied household units, 66 percent or greater.

07:37And I can slide these numbers up and down.

07:40I can change the variables to other variables if I like, or I could type in new values…

07:46…maybe I'll make this 140 instead.

07:48And I can show the results on the map based on that criteria.

07:52And immediately, I found 19 census tracks in the area that match that criteria…

07:58…and I can use that as a starting point for guiding me where I might want to look for a location.

08:04Now within Business Analyst, there's a number of ways to get that information out and do things with that information.

08:13In this case, I'm looking at the census tracks. These are the 19 tracks that match my area.

08:18I can export that to Excel for doing further analysis.

08:21I also have business search tools within the product.

08:25So if I zoom in to an area here, I'm going to drill down and start to get information on a block group level.

08:33I can also do business searches.

08:34So maybe I'm looking for, starting a Hardee's restaurant, but I'm interested to see if there's any McDonald's in the area…

08:42…and I can do that very quickly and add the selected locations there and see where all the McDonald's are in the area.

08:51So that's a very high-level overview of how you can do some market research using Business Analyst Online.

08:57Again, it's a tool for non-analysts. You don't have to be an analyst to use it. Anyone can use it.

09:04We're going to switch gears now and do some more demonstrations.

09:08This time, we're going to switch over from Business Analyst Online to Business Analyst desktop.

09:13And Kyle is going to play the role of a professional analyst.

09:17This is somebody who might have skills in statistical modeling; they maybe use SaaS or SPSS everyday.

09:24He's going to be using Business Analyst desktop to be doing some more sophisticated analyses…

09:29…and in his case, he's going to look at some scenarios where he's trying to fill in the gaps in his store network.

09:35So Kyle, over to you.

09:41Alright. OK, so, as James switches me over here, we're in the Southern California area.

09:50Does everybody in the room use Business Analyst at all? I guess, hands, maybe.

09:55OK, desktop, Business Analyst desktop? So we're kind of power users here; we're sent to learn more stuff.

10:01Was anybody in the Business GIS Summit that was on Sunday, when I gave this exact same one, if I did?

10:07I'll do something different, change the colors or something, so.

10:11But I guess what I wanted to describe today, I've got five minutes for this demo…

10:16…is a basic site selection workflow that you can use to kind of go back and kind of consider…

10:24…OK, if I came all this way, what are the kinds of things that Esri is doing for site selection.

10:29So very simple, but it's actually pretty powerful.

10:31So what we have here is a network of 169 grocery stores in the Southern California area.

10:39So we've got a fairly major player in the grocery store market that I will name nameless, but as the story goes…

10:46…our owner and founder and president, Jack Dangermond…

10:51…this was a while ago, about a year ago…

10:53…said, "We've got these guys coming tomorrow, and I want to see something…

10:58…or they want to see something on expanding a new store into their existing market."

11:02And we're just like, "OK, well, great! It's four o'clock already, and can you give us any customer data?

11:09Can you give us any analytics? What do they follow? What's their model for expanding a new store?

11:14What are the demographics? Who shops there?"

11:16And he just said, "Eight-thirty tomorrow." So we said, "Great."

11:20So the whole concept is here.

11:22We had no customer data to go off of, but we had to pick new store locations, and how do we do that?

11:27So existing locations up here, we made the assumption that if, you know…

11:32…if you're going to go to a grocery store…

11:33…you're probably going to travel maybe about 10 minutes is going to be a threshold.

11:37So we just went into the software and created 10-minute drive times…

11:42…using the Network Analyst solver in Business Analyst…

11:44…behind the scenes to get nice, detailed polygons and figure out, OK…

11:47…around each one of these stores, this is going to be a typical trade area, and what we're seeing is that…

11:56…they're actually all over the place, and there's quite a bit of market saturation.

12:01So what we did was we ran a tool, if you guys are familiar with this, it's called the Measure Cannibalization tool…

12:07…which will actually go in and figure out where you have overlapping areas.

12:10So these are areas of extreme kind of market saturation, and we figured, that's probably their MO…

12:16…is that they want to give you two sites within one area, so you have the option.

12:22Do I go across the freeway 'cause it's a little bit more convenient, or I'll just go down the way in Redlands Boulevard here…

12:28…and I'll go to the other grocery store.

12:31So we definitely have a lot of cannibalization going on here…

12:34…and what we wanted to do was we had to pick some kind of model store, so we wanted to pick one store…

12:39…that's their best producing store, and we didn't know, but actually as it turned out is, we picked the right one.

12:46And we did that running a tool called the Benchmark Report.

12:50Do you guys ever, do you guys know about the Benchmark Report tool?

12:53So, and what this did is, is…

12:56…I have an Excel spreadsheet here that I dumped out, and then for every single 10-minute drive time…

13:00…I'm comparing the entire market for variables, and you could have, you can benchmark it on your best store…

13:06…in this case, we didn't know what it was…

13:07…or you can benchmark it in the entire market average, so all stores combined.

13:12And we picked one variable that was, it's pretty remedial, but you know…

13:15…it's probably an indicator of the grocery store industry and the food industry.

13:19We picked a variable that is available with the Esri demographic data.

13:23It's total expenditures for food at home.

13:26So these are aggregated to, I guess, the block group level…

13:31…and they are the total expenditures for people buying and eating food, probably at grocery stores…

13:37…taking it home and eating it.

13:39So we found that if the market average was 100, as for an index…

13:44…James was talking about an index and if it's 100, it's average or the US average.

13:49What if you bear with me and then scroll all the way up here, we can see that the Red Hill Avenue, Tustin, store…

13:55…the 10-minute drive time around this store indexed the highest.

13:59So in this case, it was 3.5 times, or 3.5 more likely, I guess people are maybe eating food at that site, so…

14:09…different ways to explain it, but really, this is like, this is a store that we said, you know…

14:14…this is going to be our model store, and we want to look at other potential sites that will work with our market planning department…

14:21…to say, you know, go here and then pick these, and which one is going to be most like this store?

14:27And we could do that based on demographics, and I guess I'll show how to do that.

14:32So what we did, working with the fictitious market planning department at 5 o'clock on a Thursday evening…

14:38…is that we placed a series of stores.

14:42So we placed eight different stores, and then the Red Hill store is our top store, so that's the store we want to model everything else after.

14:50So if I scroll in back here, you can kind of see the methodology that we kind of used here.

14:54Again, it's pretty simple, but it's remedial.

14:57It's not in a cannibalized area; however, it is within the yellow area, which is our 10-minute drive time to something else.

15:04So in this case, if we put this store here, you're likely to have that same option of, Do I go to this store or that store?

15:12So it's kind of true to our kind of fictitious market approach here, and if I kind of scroll back over here…

15:20…I've got eight areas which I will zoom out to.

15:25And what I want to do now is pick some type of demographic profile to compare that Red Hill store to all other stores.

15:31So what I did is, I'm going in, and I don't know if you guys ever use the site, Rank Similar Sites tool…

15:39…we call it Find Similar?

15:43Well, if you bear with me, I'll just go through a few of these screens, but…

15:47…I can come through and then build some type of demographic profile here.

15:52In this case, I'm adding in variables, so if you had your own variables, like your sales per ZIP Code…

15:58…or site characteristics, maybe the number of parking spaces or the gross leasable area of your grocery store…

16:04…you add them in here, so in this case, I would have my data and some of the Esri expenditure data here…

16:10...so I've got about 50 or so variables.

16:13Now, I want to take those variables, compare them at my Red Hill store, and then see…

16:19…compared to that store, how close all of the other potential sites are going to be compared to that Red Hill store.

16:26So, again, true to the model that I was using, I ran 10-minute drive times around each of these potential sites…

16:36…and then in the end, 'cause I've baked this and took it out of the oven for this demo…

16:40…my Red Hill store, this tool…

16:44…the tool principle, the Find Similar Uses in the background, if you're statistically inclined…

16:49…principle components analysis…

16:51…so it's going to go in and it's going to look at all those statistics and all those variables…

16:54…and start to cluster like variables and factors, and then in the end, it's going to kind of mix them all together and then…

17:00…output a ranking to, say, you know, it's multispectral…

17:04…and if you try to do it on your own with all of these variables, it'd be maniacal.

17:08But principle components analysis, the algorithm within this tool, will rank them for you and spit out…

17:15…the software will spit out this layer, so we have, number 1, our Red Hill, Tustin, store…

17:20…and number 2 is our next-closest store to that main site.

17:24So in this case, it was Maywood, so compared to our top store…

17:29…is the absolute closest, statistically, demographically similar to our top store, which is Red Hill.

17:36So, and we actually had it right, so the software actually worked.

17:39It was a big win for us, and…and so…anyway, the moral of the story is we had little customer data…

17:47…and we went through a site location process that actually turned out to be, I guess, a good workflow.

18:02So this was actually pretty cool, because we had no information about sales.

18:06All we had was the location of the stores.

18:08We estimated correctly what was the best-performing store just by using Esri data.

18:13We looked at the area to see which areas might be the best-performing areas to open up a new store.

18:19It turned out, the guys in real estate planning department were already looking in that area…

18:23…to open up a store in that neighborhood.

18:26Now, the other quick last two minutes about demos…

18:28…I'm just going to show you some highlights of how you can use your own customer data with this, too.

18:33So we'll cover the highlights in this demo, I guess.

18:36Yup. Yeah, just real quick on this one.

18:39This would be the same, similar experience or similar workflow, but the concept here is that you'd have customer data.

18:46So we have organizations here that have loads of customer data, you know where your customers are coming from.

18:52So you might have a loyalty card or you can sign up for the gym, you sign your life away…

18:57…have your address on there, and they know where you live.

18:58So, just to kind of get you thinking on different ways to do this.

18:59So they know where you live, they know the types of products that you're buying…

19:02…they can start tweaking their retail mix and things like that.

19:06In this case, that's what we had here.

19:10You go to the hospital or medical facility, you're going to fill out some forms.

19:13They definitely are going to know where you live, 'cause they're going to bill you.

19:15So not to use retail but the same type situation, we want to add a new facility.

19:21And the first thing we want to do is just take a blanket look at our customer location.

19:25So we've used the Customer Setup tool in Business Analyst to figure out, OK, these patients belong to each hospital.

19:32So you start to see that spatial distribution and spatial patterns of where your customers are around each store.

19:38If I take this a step further, I've gone ahead and used what is called the Customer Drive to Areas tool…

19:44…so in this case, you can figure out where your primary and secondary customers are.

19:49So just by the locations of these…

19:51…the yellow would be the closest; 40 percent of my customers to each hospital, and then the red is 60.

19:58So these are primary and secondary areas.

20:00You might want to market to these people differently, see how many customers you have for each…

20:05…maybe how big your trade area is per…

20:09So the areas that we're actually looking for is this area, so…

20:13…these fall outside of my primary and secondary areas…

20:16…and then we want to go and mine these and figure out a little bit more information.

20:19So we want to focus on this area in this selected group of points, and we want to use those selected points…

20:27…and then we can use a tool called the Mean Store Center tool, or Find Optimal Locations is another word for it.

20:33…where actually uses K means algorithm to find clusters of points or clusters of your patients, in this case…

20:40…and then locate where ideally, based on those clusters…

20:42…are high concentrations of those folks where a facility should be located.

20:49So in this case, I'm using sheerly the location of where they live.

20:54You can run a tool in a different way where you can apply a weight.

20:57And so I'm using the same exact selected customers here…

21:00…but this is actually a location for, by the amount of billing to the patient.

21:05So just because it's weighted, you have…

21:08…just because you have a ton of customers down in the south end doesn't really mean that…

21:11…that maybe is where a new facility should go based on the types of customers that you're serving.

21:17So in this case, you would locate a facility that…definitely going to the hospital a lot.

21:23Or maybe you want to open up a rehab facility there or some type of different specialty type of medical facility.

21:31So again, very quickly, this is all I could do, all we can do in about 20 minutes, but…

21:37…hopefully it gets you thinking a little bit.

21:39If you have any questions, just let us know.