Art Institute of Chicago (ARTIC) - The Art of Analytics - Inspire 2017

Museums are complex organizations, although most people don't realize or recognize this fact. Most museums combine a number of business models within one operation and consider both mission- and financially driven objectives in everyday decision-making. What's more, museums face significant headwinds, such as increasing costs and an uncertain economic environment. In this session, you'll hear how the Art Institute of Chicago (ARTIC) has leveraged Alteryx to grow attendance and operating revenue through more informed decision-making, strengthened operations to share more of its collection online, and understand its audiences with the aid of machine learning. Through specific examples of the museum's successes, failures, and efforts in-between, you'll gain relevant insight into how to introduce or scale analytics in a resource-limited environment.



Video Transcription


Andrew Simnick:
Hi everybody, I'm Andy from the Art Institute of Chicago. My colleague could not be here. He's Matt [Norris 00:00:08]. You'll see him running around, actually smart enough to go to a different session right now, so sorry everybody. [Ramon 00:00:19] said, I oversee finance, strategy, and operations for the museum. Who here has been to the Art Institute before? All right. This is good. We find that it's a lot more fun presenting from the position of an art museum than a management consulting firm. I'm glad you guys enjoyed it hopefully.

What we start with is, what is the art institute? You think on its face it's a venerable institution. We're well over 100 years old. We're an art museum in Chicago, but it's really spiking on all these fronts. We were founded in 1879 with the same mission as today, to collect, preserve, and interpret works from all time and all geographies, but we're known for collecting works of today so we've always collected what was contemporary at the time. Our impressionism collection was contemporary at the time. Our modern collection was contemporary at the time and this is becoming more and more difficult as the art market gets more and more complex and expensive.

We're known for quality and scholarship. We have a saying, we don't know when to quit at the art institute, but it's what's given us the reputation we've built over many years and we have a fully functioning art school, the School of the Art Institute, within our same corporate structure, which is unique amongst the major museums in the US. Also, has a ton of complexity, but it's great for the business.

By the numbers, we're about 1.5 to 1.7 million visitors per year. This has been on the rise. We've grown, from an admissions revenue standpoint, about 20% year over year over the last five years, which we're very proud of. We're super prolific. We do about 30 exhibitions a year ranging from large shows such as last year's Van Gough's bedrooms to smaller department level shows using the permanent collection. Each one of these has an importance whether to visitors to growing attendance to expanding the art cannon. One thing we're very proud of is we've been voted a top three museum in the world in Trip Advisor the last three years. We're also gigantic. We've about a million square feet. That's puts us at number two, maybe number three depending on how you define it of museums in the US. That adds, again, a ton of complexity particularly given the value and importance of the objects that we have inside.

We have art. As I said, we're known for collecting in the moment and that's allowed us to get some of the icons. This one is pretty famous. Most folks know it from [Ferris Bueller's 00:02:39] or if you grew up... Ferris Bueller's Day Off or if you grew up in Chicago you usually saw this in the text book in third grade, but it's awesome. Just be in the museum. Night Hawks is at the Art Institute. American gothic will be at the Art Institute soon. It's been out on loan, it's first trip to Europe as long as it's been in existence. Then you see the target. We recently received our biggest gift of art in museum's history, which was a series of over 40 contemporary icons, which allows us to keep going along the strategy of having iconic works collected in the moment just in a way that's compatible with the market.

Question we always get, so you're an art museum, why do you need data? You're are in art. That's qualitative. It's static. I couldn't disagree with that more. We're complex. You think about what a museum is, at its core we're three things. We are a collection. We are the people and we're the building and the place to contain it. We consider that the core of the museum, the collection is especially. With the collections we're the Art Institute, without it we're a really poorly designed aging facility on expensive real estate in Chicago.

If you think about what we produce, it's bunch of different things. We have a visitor experience, somebody comes in, they buy the ticket and see icons from around the world. We have an educational program. Over a hundred thousand school kids come to the museum each year and has is a different experience than someone who buys a ticket and comes through. I mention the exhibition of all shapes and sizes and topics and the scholarly research that appeals to admittedly, a narrow segment of the population, but just as important for us to carry out our mission. Then you think about the audiences that receives this. We have students. We have patrons. We have visitors. We have collectors, scholars, and this is not a full list.

What happens is all of this goes back to the collection but there's different business models underpinning each one of these with different metrics, both financial and non-financial and sometimes conflicting. We deal with data. We view it as a way to demystify the Art Institute to isolate different parts of our business model and then make it as simple and as efficient as possible to then allow us to put more resources, both time and money to supporting our collections, and ultimately, our mission. That's what we do.

One thing that makes us different, we are a behemoth amongst museums. We are tiny compared to all the fortune 500 companies. We have no single metric for success. Sometimes it's money. Sometimes it's feet in the door. Sometimes it's subjective quality and that's okay. It's just part of how we work. We do have significant resource constraints in the world of analytics. Our team right now, I would say we're between two to three FTE. We’ve one dedicated person and then we steal folk's time where we can as well as pretty lean infrastructure setup. We, as an institution, are limited in analytics. We have some of the best people in the world at what they do in terms of art history, in terms of visual interpretation. We're not a very quantitative place. We do have pockets of excellence, but we're better at raw horse power and intelligence.

We have fragmented systems. Everybody has legacy systems. I will admit I still use Excel from time to time. It's okay. It's not for everything, but we deal with a mix of best in class, state-of-the-art databases all the way to home built systems from the 80s that require custom extraction. It's tough. You have to get on one spot.

Where does Alteryx come in? It allows us to scale up. We have tons of different tools. We use Microsoft SQL server to warehouse and distribute static reports. We use Tableau for visualization, but we put Alteryx right at the core of this. What this allows us to do is get the leverage we could get with more FTEs or more investment. Instead, we have really talented folks with the right tools that can take data form any of these internal data sources, ticketing, membership, you can read the list,  pull it into SQL server, use Alteryx to automate our workflows and to allow one person to oversee quality, prioritization and also pull in external data sets. That's some of the examples of what we use. We visualize in Tableau to capitalize on the strengths of the organization, the visual analysis and the formulation of questions, and then we output visually where we can. We also have static reports, excel extracts. Whatever the best tool is for the job, Let's do it in an automated, lean, consistent way.

I was planning on doing today is talking through some of the biggest use cases and business questions we've answered over the last couple years. I would also say I'm also a kind of questions in the moment guy. That might be against Alteryx's rules, but if you have any questions interrupt me. It's okay.

First thing we got to is attendance. When I joined the museum the way we had treated attendance was very monolithic. We have a number of people in the door, we have [inaudible 00:07:58] 1.5 million, doesn't matter where they're from, what time of year they come, what channel they come through, just feet in the door is what matters. We didn't think this was right. Something seemed a bit amiss. We started asking questions. What actually drives attendance? Is it exhibitions? Is it tourism? Is it price, programming, weather?

We get a mix of tourists and local visits. We're about 50/50. Paid admission, 75% tourists, 25% local. Members, 90% local, 10% tourists, then we have ton of free admission channels for local audiences, but here one of the first things we did was just draw our attendance patterns by zip on a map. Then, look, we have international audiences. We have visitors from all 50 states and then, when you get into Chicago you definitely see density stretching along. The Art Institute is right around there. You see it stretching along the lake going up through the north west and then density much higher in the city than the suburbs, which came out to be an important insight, which I'll get to.

We start with data like this. This is hard to forecast. This is hard to interpret. This is three years of data. We have lots of peaks and valleys that seem somewhat tied to exhibitions, which are the beige shaded areas, somewhat tied to seasonality, some peaks in the summer, but ultimately, you would just look at this and say, it's up and down but it's kind of flat. We've got some thumb growth in the tail end, but it's hard to pinpoint what actually moves the needle. For a research constrained organization, that's not a good place to be.

The first thing we do is break it up by geography. We ran some cluster analysis, saw which groups behaved similarly. Interestingly enough, this was our first jaw dropping moment. We saw that visitors from the city of Chicago behaved differently than visitors from the suburbs of Chicago, which immediately told us we could shift our marketing spend more into the city to get more return, which was easy enough to do, made everyone's lives easier, reduced complexity, helped us grow local attendance. What we also found is tourist [that's 00:10:10] seasonal, independent of exhibitions. We rise and fall with tourism to Chicago. This is just US fly market.

When you pull it out it becomes much cleaner. You just see a nice [inaudible 00:10:20]. When we communicate this internally we didn't come at it from our P value is X, this is the scatter plot that shows the correlation. We just said, hey, look. There's a pattern guys. It's up and down and it's independent of these beige rectangles and I was able to get folks attention. Member attendance, interestingly enough, we always had the hypothesis that it was mostly exhibition driven, and here we felt pretty confident just by separating it out by channel. Given it's predominantly local, we put them all in one group, but you see these table tops where you have an increase when there's an exhibition, pretty flat otherwise. What you see here, this was our Van Gough's bedroom exhibition I talked about earlier. This was our biggest show in the last 10 years. That really moved the needle which allowed us to get over 100,000 members.

What's happened with this is we've become much more cognizant of what events we can put on to drive marginal member attendance. Again, just to break it out into comprehensible components, to break that initial view of all visitors are the same, but that's not enough. Just looking at it doesn't allow you to make business decisions on the fly or see if something's changed. The first big model we built internally was multivariate regression model by region and by channel, city of Chicago, suburbs, drive markets, fly markets, and international for tickets and members. Then, we took data from inside, the factors we can control. We looked at ticketing data, membership data and other random Excel files, translate PDFs around the institution as well as factors outside of our control, things like Chicago tourism, the city events schedule, like Lollapalooza, it's right in our backyard, or weather data, Chicago, you might have heard, has really erratic weather.

By doing that we then came up with this, which I'd say is our first real institutional dashboard that's still exists today. What we show in the top is actual versus expected attendance by week. Expected is what the model tells us we should expect based on the factors that matter. Actual is what actually happened. We have a couple different versions of this. This one aggregates local versus tourists and at the bottom we stole a trick from a manufacturing reporting and put it in a control chart. Our organization glosses over when you mention statistics. No one really cares. They just want to know the answer and move on to the interpretation.

What we did is design this thing at the bottom where if it's a solid circle, business as usual. If it's not solid, something happened. If it's above, something good happened. If it's below, something bad happened, but it reduced our meeting time from a couple hours a week to having one every eight weeks with folks having a better idea of what's going on. About three months after we launched this into the wild we received the number one Trip Advisor museum in the world. We didn't know what to do with it. That's great. Our visitors like us. It's a vanity metric, what could it do to help us? We saw a break in our model specifically for local ticket attendance right here and we had no idea, but we noticed there was a lot of press coverage. The Tribune, the Sun-Times, a lot of TV coverage and national publications picked it up as well, but we didn't see the same lift happening in tourist groups, in membership. It was this local ticketed attendance.

Internally, we made the decision, truly strategic, to shift our marketing message to the number one rating versus an exhibition the following year. Once we got the tourism season we saw this. This factor has persisted in our model since we've... In the first year since just making that decision to shift with seeing the lift happen concurrently with the launch of that ad campaign. In the first year alone we increased our net revenue by about two million dollars, which for a museum with a 100 million dollar budget is a real big deal with no additional work. This was the first real use case where the museum as a whole really got behind the idea of analytics. It wasn't rocket science. We're not great at this stuff. It was just communicating something in a simple way that showed meaningful impact to the folks that cared, which in this case is the entire institution. Visitors plus money is a good thing.

What we've since done, since attendance is the driver for much our business, we've taken this model and translated it to other parts of our business, so forecasting museum shop sales. The one I'm personally excited about it financial forecasting. We now let a computer tell us our budget projections couple years out and this year it's working pretty well. Membership joins and renewals, and then frontline staffing, which is another application we hadn't thought about when going after the first step. We've built something here, scrappy, lean but it works, and it's translatable to a lot of other business we have, which allows us to get maximum bang for our analytics investment, both time and money.

Just to show you the shop model, this is an actual versus expected. We've gone through a huge effort to reduce skews, cut back on inventory, scale back our shops. We've seen in using this tool, we've been able to do that without impacting top line revenue, which still gives us financial impact. It's less work, less complexity, and we didn't have to do anything else. We just rebuilt the model and used the same view that everyone's been familiar with. Again, this is as much leverage as one can get.

What has been great is being able to thread analytics through the revenue generating side of the business, which is similar to what many of you may be doing in your organizations, but we're not just revenue generating. We have other aspects of what we do. One of which is serving a diverse population. Not only do we look at generating paid attendance, increasing membership, but also how do we introduce the arts to new audiences. How well are we doing with community engagement? Who comes to lectures and performances? The challenge with that is traditional market segmentation research just doesn't work for an organization like ours. Traditional has been very manual, driven on surveys. You do some analysis on census data, but that doesn't give you a lot of psycho-demographic information.

To really be actionable for an organization like ours that is complex as it is, we started with Chicagoland. Several years ago we ran Chicagoland market segmentation. Lo and behold, we do well with audiences that like visual arts and we don't do well with audiences that don't like visual arts, which was very insightful, but it also wasn't that actionable. It's a good thing for us to know, that we don't appeal to everybody, but where do you go with that? Looking at what we call the museum segmentations, so where does the art institute fit on that spectrum of interest in visual arts. There's where we have to start thinking about is it worth the effort to get that information, relative to knowing who likes visual arts. Then, to really get down to program segmentation where someone coming in as a tourist to see American gothic may be a different profile than a connoisseur coming in to look at our new contemporary acquisition.

For us to really decouple and understand parts of our business model we have to get to that level. Alteryx allowed us to do that by eliminating the resource requirement. We've always had this concept. We've never had the tools until very recently and what we've done is used a combination of the geolocation tools in Alteryx as well as machine learning. Again for us to be doing machine learning in an art museum it's not that hard to do using the tools. By using a combination of this, what we've tried to do is... We call it a programmatic segmentation approach where if we know what types of audiences are coming already can we train in the algorithm to figure out where more folks of that demographic and psychographic profile exist within the Chicago land area?

We think this is pretty universal. We can apply it to anything. We tested membership first. We had a decent sample size. It has both mission and financial components, and it's data we're very comfortable with, more robust sets. We picked it as our first use case. We first identified high penetration census blocks within Chicagoland. That covers, I believe, 90% of the member base despite being much less than 90% of the geography of Chicago. We then use machine learning, figure out what are the different factors using some of those external data sets that correlate well to areas with high member penetration, and then we applied this schema to the entire area of Chicago and Chicagoland to say, where are we underperforming relative to expectations?

Another 'aha' moment for us along the way was this idea of [sentences 00:19:30] block versus zip code. We originally thought we were really good because we could do the zip codes. Chicago's zip code doesn't work. It works in certain areas, but around the edges of the city it gets a little wild. What you see here is the zip code covering Oak Park and Berwyn. What you see with the red, those are areas that correspond incredibly strongly with membership. We have some areas of blue, which to us, represent opportunity, then the gray areas where it's the cost of member acquisition is going to be much higher. If we have one dollar to spend or one hour to spend we're going to go to a red or blue area.

By getting down the census block we can focus our energy even further to then make our time and our money go that much farther to grow our member base, yet, allowing us to do what any business would do except in a much leaner and a much resource constrained way. The exciting part, which gets me really fired up about this, is it's applicable to all of our channels. As long as we have some basic data of where people are coming from we can do this. E-commerce, we're currently undergoing a project to really supercharge our ability to sell tickets, memberships and shop product online. This allows us to really understand where are additional targets, and the richness of the data is just extreme relative to what I've had in the past.

Lectures and performances, we put on a very robust schedule and we understand that it's also a niche audience, but if we can identify those niche audiences we could effectively market what, admittedly, has been resource constraining before. School tours, which is super exciting, where can we reach out to Chicago to get more school kids in the door for field trips and get early exposure to art and community outreach. Are there neighborhoods that we have a chance of increasing interest in the arts, making a name for our self and also putting the Art Institute into areas where, again, it helps us fulfill the mission, and get more folks engaged with visual arts where we have a chance to engage them with visual arts.

Go from that into the collection itself. This was our more unique use case. Given the success we've had on visitors, both for revenue and mission, we shift over to think about how do we manage the objects themselves. A couple things magically worked for us along the way. When we started doing our analytics work the museum was undergoing a large scale effort to get more of its collection online, so taking high-res photos of every object going into the galleries to prepare for distribution on our website or other digital platforms as well as just thinking about ways to be more efficient about how we handle collection related operations.

Mentioned our home-built systems. This is our tank. Every organization has that one database that it's like... I used to drive a '98 Camry until very recently. This is our '98 Camry, except it's older. That's our database. It's home-built. It works incredibly well, but it's meant to store data, not to report on data, which gives it a pretty nasty reputation within the museum. We have this setup with SQL Server, Alteryx and Tableau. All we did was build an export function to take it from a database format called 4D, away from the front end that isn't great reporting into something that only does reporting. I can't [comment on 00:23:02] how much time the team spent but I think we spent in meetings than actually making that link happen. What it's allowed us to do is come up with visual dashboards for information in our collection itself, which I think is the simplest thing in this deck, but it's also been the most transformative from how we do business.

First, we're now up to 98% to 99% of the works in our galleries online with an image. This is coming up from a base of less than 50% three years ago. Once we put this tool into effect we got two to three expert activity with the same number of staff just by allowing our handlers and photographers to identify packets in the gallery that needed to be shot or in advance, seeing rotations and being able to catch objects on the fly. You see some red space here. I have to talk to some departments when I get back to Chicago, but again, we're in a much better spot and prepared to disseminate information through digital channels.

We've also learned more about our collections. Our trustees and our support groups have really gotten behind this idea of using applied data for decision making. We wanted to see if we could learn something about the growth of the collection and we could. All you see here is a cumulative sum of objects in our modern and contemporary collection shaded by media. What's fascinating here is you see this pattern where each time we introduce a new media you see very slow growth but then expansion. As we think about in today's world with digital art this type of visual now has us thinking about what investments do we need to make from a infrastructure standpoint, a space standpoint, a talent standpoint to keep up with the growth that we now project in what we call time-based media, audio-video, computer-based art. It's just something new. If you look at the patterns in the past, for instance we have installation art or photography at some point we didn't have those. Now, it's a major part of the collection.

We've done this for every department in the museum. We've identified areas of overlap, areas of need. Again, all it is aggregating data that we already had, just communicating it in a new way and to an audience that is receptive but also just interested in knowing this information. This is my favorite example of the entire document. Data, for data sake is cool and it's exciting, but it doesn't capture imaginations.

We have a really great digital design team led by a gentleman named Michael Neault. He partnered with our education department headed by my colleague Jackie Terrassa to come up with a... We call this a platform for bespoke navigation. It was an idea of how can we use the new dataset that we have to help families unfamiliar with art go around the museum with parents being empowered to explain something to their kids and the kids being engaged. We came up with this platform called Journey Maker. It actually just won an award itself by South West, which we're really proud of. We think it's a tool that we've now created that could be really expanded for what we call bespoke navigation, which is really awesome. I'll just let this play.

All this platform is taking clean object data, images, location data and an algorithm, giving it a spectacular design interface and now we have families running around museums with maps just being super excited, whereas, even just two years ago, they may not set foot in the building. That's where it really comes together for us.

Just thinking back. We've been at this for about three to four years. We've definitely had our fair share of wins, losses and everything in between, but just consistently with the initiatives that have really taken off an changed the way we do business. First definitely, play to your strengths. We are a visual culture. We're not quantitatively oriented, but we're just insanely smart and curious and can ask really good questions. We really based it around prioritizing questions first, communicating in a visual way and doing it in a way that resonates well within the art institute, may not be [compatible with 00:28:27] all organizations but it's what's worked for us.

We only pick questions that matter. We don't have time. We don't have money and we only have so many shots on goal before the resources are needed for something else, so it's always critical for us to just make sure we're doing what matters most in the business. Lean processes or bust, we can't afford to do anything overly complicated. It's got to be simple. It's got to be quick. Also, touching on that bottom bullet, good enough is what we aim for when we're launching something new.

Then I think the part that is forgotten quite a bit and when we've succeeded and when we haven't succeeded we haven't is the personal element. At the end of the day you're helping somebody make a decision. You're helping a business leader take a risk and do something different or you yourself are selling an insight that your company or industry hasn't seen before. That's all about the personal engagement, interaction and making your colleagues and your teams feel comfortable with taking a leap. I think, thankfully, we've had enough of these where we've taken enough leaps over the last couple years to really change the way we think about how the Art institute works. With that, thank you for the time. I know there's some strong competition. I didn't have free T-shirts for you guys, but I'm happy to answer any questions that you have.

Crowd question:
Does the museum currently collect any data around patterns of traffic throughout the building and are you leveraging that yet?

Andrew Simnick:
Yes, and yes. Similar to resource constraints around the segmentation work we have migrated from having observers take tally marks of folks going through different doors to using aggregated Wi-Fi signal to heat map the museum. While we don't do individual level tracking, it's all anonymous and aggregated it's allowed us to understand things like, for instance, we used to design an experience based on the entrance. If you come in on Monroe Street by the modern wing it's different than coming in Michigan, which is the one with the lions.

We see that there's a super highway through the museum, no matter what entrance you come into. The dual time and probability of visitation is within 2% for any gallery in the museum. That's a lot of to simplify things and have one approach. We also can see higher dwell times in certain galleries that are either icon rich or adjacent to areas of high traffic. One thing we challenge ourselves is how can we break some of the patterns to get higher traffic to less visited parts of the museum, but understanding, just from architecture and visitor preference there are some things we have to work around. Because of that data we've been able to cut a lot of the manual work as well as understand a little better how folks use the space we have.

Crowd question:
Thank you.

Crowd question:
You mentioned that your analytics team is pretty lean. Can you describe the composition of our team and how you're organized?

Andrew Simnick:
In our original inception, I came in as the head of strategy. Half my time was dedicated to this type of work and then I met Matt, who was working in our IS team and responsible for connectivity between the enterprise systems and we're able to take half of his time too. I'd say we had one FTE at the very beginning when we did that attendance work as well as porting over, drawing that initial link from our collections database into the SQL Server.

Matt now does analytics full-time and my role's changed a bit where I don't focus on this as much as I used to. What we've been able to do is we have a hub and spoke model where Matt acts as the oracle of analytics and report on folks responsible for reporting or just a curiosity of business will come to Matt. We have an informal reporting relationship and that's where it ebbs and flows to how much resource we're dedicating in the museum to a project at any given time.

While we haven't really gotten past that one FTE dedicated in earnest, it really ebbs and flows based on the [business 00:32:45] of the day. I think it's also given us the flexibility and helped us with the buy in that if we went to something super-centralized right off the bat we might not be able to get it. It's a great question. We wrestle with it all the time.

Crowd question:
What predictive tools were you using for your forecasting models and how would you weight for your various exhibitions in that?

Andrew Simnick:
For the predictive forecasting now, we set our exhibition schedule out three to five years in advance for the major shows. We do have some scoring based on the likelihood of popularity or artist recognition and put those into the model. We do have to take a guess at weather patterns, which it screws up, we don't always get it right, as well as rely on tourism projection from the city with some adjustments we'll make internally based on how much we believe chose Chicago in those numbers. So far this has been our first fully year of being predictive with our attendance projections. Through the month of April we were off by $5000 on a base of 13 million. We feel pretty good about it.

Again, for most of the year we saw minimal variation. We saw one period where we're low, which corresponded to a 70 degree weekend in February around President's Day, which really walloped us, but then also unexpected growth through our new medieval gallery installation, which we didn't think would have a lift, but since we'd shift our marketing campaign to that it's our first permanent collection installation that's just showing a lift. Those two offset by the end of the year [inaudible 00:34:21] up, but it really is just going back looking at just normal seasonality, our exhibition schedule and then hoping we get Chicago tourism right. We did linear regression and we do it by week too. If we do it by day or by month it get all out of whack.

Crowd question:
Any other questions from audience? I have one quick question. It's obvious that you adjusted how you rolled out your dashboard so your team, at large, would be interested. Do you have any advice, tips and tricks for the audience in terms of really getting that buy-in when it comes to creating the culture of analytics where one may not naturally exist?

Andrew Simnick:
Yeah, I definitely have opinions on this. Rapid prototyping, the combination of Alteryx and Tableau is amazing for building a tool in collaboration with someone who is new to analytics. For all of the success we've had it's getting the right people in the room at the right time and then just sitting down, what do you think about this. We'll drag it in or we'll change the graph. I'm allergic to pie charts, but some people like them. It's okay, but also, I say a Tableau report you can build... It's almost like a super charged pivot table and that works too. What we've done is taken the flexibility using a rapid prototyping approach with the end user in the room then taking that and figuring out what's the best channel.

Immediately, not everything should be visualized. We use automated reporting out of reporting services all the time. Typically, we'll take someone's prototype, hard code it and then it just gets dumped in with normal email reporting to the level of convenience and phone readability that a lot of our teams expect. Besides using the tools it's also the communication. [Whenever 00:36:21] it's personal as long as somebody wants to use it, feels good about it, they'll use it. If they feel like it's being foisted upon them and they didn't have a hand in the design they'll just languish.

Crowd question:
Thank you for that.

Crowd question:
I understand your background is McKenzie before you came to Art Institute. With that background did you feel that you knew what you were looking for when you started on the analytics projects or was it a case of a lot of things appeared magically? New unknowns.

Andrew Simnick:
It's an interesting question. My background before McKenzie... I was actually a research scientist for a while, which is a real natural evolution to arts and culture. I think what's been good for us is asking questions. Every step along the way whether you forming an experiment for a publication or at McKenzie where everything is amorphous until you get to know the client and can figure out what the problem is, and then even now, I think this idea of taking a hypothesis-based approach, framing a question early... It's helped us get closer to that right first question before we do any analytics.

Admittedly the approach we use, it's so new. We weren't doing this at McKenzie when I was still there. Tableau was just starting to pick up a little bit. I think what's been good is that approach of just asking the question, framing right off the bat. That's time-tested. It's been around forever. These are just new tools to test that question a lot faster. Folks getting [in and out 00:37:50] today, I don't think there's that much... The experience curve doesn't really matter because the tools are changing so quickly. That answers the question.

Crowd question:
Any other final questions?

Crowd question:
What would be your recommendation to the analytic community who's trying to work with the arts community to try to do what you've done at Chicago?

Andrew Simnick:
I think a lot of it has to do with finding the right use case. The attendance modeling in our previous administration, both the COO and our VPO marketing were very much in favor of getting it locked down and getting a better approach to growing our audiences. At that point, finding the right leader in the senior tier and then moving from there. With the collections, it was finding the right use case. Our curators, admittedly, not as impressed with growing admissions revenue, they're happy, they're proud of it, but it's not core. Being able to apply the same tools to understanding the collection, super charge buy-in and it's finding those winds in the right pockets to get an audience.

I think what's interesting is, and I think about it, there's a lot of similarities between arts and healthcare, in that you have an audience that is new to these tools, [there's 00:39:13] an ingrained or an institutional approach. I think with arts though, everybody's naturally curios. As long as the tool reflects what matters most to the curator or to the conservator you got a real good shot at getting it across the goal line  because they want to improve. It's just finding that thing that somebody really cares about and that opens the door to other conversations using data.

Speaker:
Thank you for that, and with that I would like to thank Andrew for a fantastic presentation. Before our audience leaves today, if you could pull out your phones and rate the session that would be much appreciated so that we can continue to improve your experience here at Inspire. Thank you.

^Top

Experience the
Power of Alteryx
For Yourself.

Get Started