DAVID LEDGERWOOD: Thanks for joining us. Good to have you on today.
JASON KIESEL: It’s great to be here. Thanks for having me.
LEDGE: Would you mind just giving a little background story? A couple of minutes about yourself and your work so the audience can get to know you a little bit.
JASON: Sure. I am currently the founder and CEO of CitySourced. We provide software for local government agencies to connect back-office systems with citizens. We kind of created the pothole app, so to speak, way back in the day – back in 2009 on stage at TechCrunch.
I got my start in technology way back in 1997 after graduating UCLA. I started out as a web designer. This was back when all your design was done with tables and there wasn’t really any CSS or anything like that. It’s great to have been in a space so long where you’ve seen such a dramatic evolution and change.
At the same time, it’s much easier to start companies and write software now but there’s a lot more competition, there’s a lot more stuff out there, so in general it’s a great space to be in.
JASON: Oh, yeah. I ran a hosting business.
Tell me about civic software. I mean, gosh, selling to government. Yeah, that’s a peach. Been there, done that. We know a lot of people that want to make government better and entrepreneurs that are interested in upgrading the technology of the government. I bet you can tell some stories about that distribution process.
JASON: Well, selling to local government, the IT space for local government is a $60 billion a year industry. That includes software and hardware. It’s not surprising that a lot of the folks we deal with still have on-premise software solutions.
You’re starting to see a migration to the cloud. We do occasionally get requests whether or not we support our own software on-premise, and the answer there is typically no. But it’s interesting in the sense that, from a procurement standpoint, that’s where a lot of people push back and say, “Oh, my gosh. Selling to the government is so hard.” It is. I’m not going to lie to your listeners and say that it’s not. It is more difficult than selling a $10 service that you sign up for, freemium model type of scenario.
But the added benefit is your churn is very, very low. Once you get a customer, once you’ve acquired one, they tend not to leave so long as you can fulfill the service and make them happy. I think our trailing 12-month churn is 99.6%, something like that. We retain 99.6% of our customers. Very, very little churn over the long haul.
Our sales cycle…
LEDGE: All the SaaS founders are like, “Hey, 0.4% churn. I’d love to have that.” You're right. That procurement thing is such a big deal. You probably have to wade through RPs and different things.
Are you able to leverage the existing relationships where you can say, hey… I know from my experience selling to states that it’s really much, much better once you have one or two because then you’re at the table. It got easier and easier to win when you had those proof cases because everybody wants experience.
JASON: Definitely reference customers are important. It is a small community. There’s a lot of kind of conferences specific to GovTech, or the pub sector. You’ve got like NACo, the National Association of Counties. You’ve got ICMA, International City Managers Association. You’ve got these little groups, these pockets of constituents that all share similar titles, and they share stories and best practices. So, having reference customers is very, very important.
At the end of the day, having a quality product. You can’t really run anything without that. Folks are going to see through that very, very quickly.
LEDGE: Absolutely. There’s quite a bit of scale that would go into such a thing.
Tell the story a little bit of, how and why you jumped into this and what’s the story been like scaling up and all that business?
JASON: It’s an interesting story. I started CitySourced as a side project. The initial business that I wanted to start was essentially grassroots activism. I did a three-year stint at MySpace. I was a contractor/consultant for them for three years. Built out their IM client with a small team, and later transitioned to just some general product within MySpace.
But, having been in that environment and seeing the power of the social network. I was super into politics – local politics. Listening to talk radio on my way to work, et cetera. I wanted to build a social network that was strictly geared toward politics. I wanted it to be non-partisan.
So, that’s what I did. I built out a non-partisan political social network where folks could discuss issues. I feel like people were a little more civil back then. This is back in 2007/2008. It’s now progressed into just a full-on flame war on Facebook, but back then it was still fun.
Actually, what I had done no one had done ever. You could always go online and find your federal representative or your congressman, your senator, et cetera. You could go to the state website and find your assemblyman, but to find everybody in one spot was difficult. So that’s what I did. I went through and scrapped every single politician on the local, state and federal level and put them all into a single database and then essentially wrapped geocoding services around it.
You could enter your address and we could figure out who your councilman was, your mayor, all the way up the chain to president – who at the time I think was George Bush.
LEDGE: People who are maybe listening, some of our younger listeners, like, “Big deal, man. Every site has that.” This was a tremendous undertaking at that time.
JASON: Yeah. We had about 60,000 politicians in our database. It was the largest one that I had found or knew about out there.
LEDGE: There were no APIs even. Nothing.
JASON: No, no. It’s not that long ago, but it’s amazing how much we have progressed. But yeah, there was nothing like this.
That was my product. You came to the site and we would tell you who represented you and we offered some form letters that you could write out and check the box; I want to send this to my mayor, my councilman, all the way up. You could check them all. You could share that with your friends and so they could come and sign on to your letter and send it to their reps. We thought about that viral interaction.
We actually got a couple of politicians to send us cease and desist letters because we were sending them so many – because we actually used eFax at the time, which was an old school hack of flooding their fax machine which is printing out these letters of folks wanting to do whatever it is they wanted to do. We would get faxes back saying, please stop.
It was a lot of fun.
So I had this database of 60,000 politicians. It was right around when the iPhone had just come out, and I wanted to build an iPhone app. I didn’t want to just build a game or something like that. I had this asset that I wanted to leverage.
The story is that, how could I leverage this database? I thought, you know, there’s stuff all around my neighborhood. Why don’t I just snap a photo of it, I’ll use the GPS on the phone, and we’ll just send it to the politician. He’s the one that’s responsible.
Well, lo and behold, the politician is not responsible for that. He or she makes laws and legislates, but it’s actually like a city manager or a department of public works that does that. I didn’t know that at the time but, whatever, I had this large database and I wanted to go do something so that’s what I did.
I created a proof of concept iPhone app. I went onstage at TechCrunch 50 in 2009 – September 15th, 2009 – presented and ended up getting some pretty good accolades from the crowd. We were kind of this unknown… I actually got in because my childhood friend is the executive producer of the show, and I showed him the app at a party and he’s like, “Oh my gosh, we’ve got to get this into the show.”
We only had three weeks to prepare. It was kind of this whirlwind of like everything happening correctly at the right time, and that’s how we launched CitySourced. It was a lot of fun.
LEDGE: How has it evolved since then? Even that is a long time. You’ve got 10 years there.
JASON: Yeah. That was a long time ago.
LEDGE: What have you done and what have you learned? Assumptions that were amazing and assumptions that were not. Those are some of the best stories.
JASON: The big assumption early on was that politicians were the ones that mattered, when in fact they don’t when it comes to both a buying decision... So if you’re selling a product, they weren’t the ones making the buy decision. Sometimes they were, but that could actually backfire on you.
We had a couple of instances where our product was procured by a councilman and then that councilman lost his seat, and because there was no other champion within the organization especially in the back-office – public works, IT, communications, etc – we lost the contract. That was an interesting learning.
We actually learned that very early on, was that we needed to be embedded in the organization, the city infrastructure so to speak or all the various department. We need to be embedded in those other departments and find champions in those spaces early on.
From a sales cycle standpoint, I’d mentioned earlier the procurement process. We’d go to hit up VCs for money for capital to grow the business and the first thing you say is, “We work in government,” and they’d be like, “Not interested.” Which is interesting because there’s a lot of money spent on IT and software in government. They just didn’t like the sales cycle. They thought, oh, it’s going to take you 18 months to get a sale. Actually, initially the sales cycle was only about two to three months. It wasn’t that long.
Now, as we pushed upmarket, kind of made our product a little bit more robust, more enterprise as we would say, the sales cycles have gone up. I think our average is about anywhere from three to six months now, so it’s gone up a little bit. Those sales cycles are anything from RFPs to just forging relationship from going to trade shows, et cetera.
That was something we learned early on. The money guys, the institutional money guys, the VCs, they were telling us one thing but we weren’t seeing that in reality. I think their experience in public sector was – and it still is for… You take a seven-figured project like a massive ERP implementation that’s $10 million, $2 million… I think LA implemented a CRM a few years back and it was $16 million. That is going to take 18 months – 18 to 24 months. There’s a lot of hoops you’ve got to jump through. But for software that our average contract is 25,000 a year, for that a lot of times you fall before RFP threshold, you fall below certain things.
That all varies. There is no standard RFP threshold. Every city has their own legislative process for it, which I thought was really interesting as well. We would get an RFP, we would respond, and as part of the response they would want us to… First of all, you’d have to print out the RFP five times, stick it in five separate envelopes. Then the digital versions, you’d have to put the files on a CD-ROM and put that in the envelop as well.
LEDGE: Right. Which had to be Word files.
LEDGE: I've done this. So yes. I’m smiling because, yes, I've been through this. Print out a truckload of binders to bring them in.
JASON: Lots of trips to Kinkos and The UPS Store, but it pays off. You win the deal, you get that customer. I think our average lifetime with the customer is seven or eight years.
LEDGE: How has the product evolved? I know you talk about the pothole app, but you can request a lot of different services and needs, right, through the application?
JASON: We went really deep. It’s called service requests in the space. That’s what they generalize it as, and sometimes refer to it as 311 for those cities that do have an actual number. Think of it as 911 for non-emergency issues.
We went pretty deep on that. Creating workflows around what happens if something stays in a status a certain number of period? Well, you’ve got to escalate it up and do all sorts of stuff.
We also built out our stable of back-office integration. A lot of cities run on various software. Things like Cityworks, Lucity, Accela. Things to help them manage their assets. Ideally these things are called asset management systems. So, where are all my fire hydrants? Where are all my street lights? All these things that no one ever cares about.
LEDGE: Stuff that none of us ever think about.
JASON: It doesn’t matter. Essentially what we’ve done is, we’ve taken the request on the citizen side, associate it with an asset on the city side, created a service request sometimes creating work orders, it kind of depends on the system. But, once that data is in that back-office system we then continue to follow it and then update the citizen when updates are made in that back-office system.
We don’t disrupt the workflow of the city worker, which is ideal. There’s no training involved.
LEDGE: That’s critical. For anybody not paying attention to that, that adoption process. That works, and ought to work, for any industry. So if you’re trying to approach somebody, do not screw up their workflow. Find a way to make your thing plug in.
That was not common knowledge, it still isn’t common knowledge and it ought to be. Software fails because of adoption, not because of features.
JASON: Yeah. It’s very important in the public sector. Most of the workers are union. A lot of them have labor contracts that, for better or worse, say that they’re not required to learn new software, they’re not required to learn new practices. They get trained on a certain piece of software and, for all intents and purposes they expect that software to be somewhat the same throughout a long period of time.
LEDGE: For 35 years.
JASON: As a startup you’re like, let’s iterate, let’s build and push out product all the time. We would actually get pushback from our customers saying, “Hey, you guys just pushed out a new release. The button went from the right side of the screen to the left side of the screen. I now have to train 300 of my call center guys that the button moved.” We’re like, well, the button just moved. Can’t you just tell them, “Hey, the button’s not there it’s over here?” They’re like, “No, no. We have very rigid, structured training processes. These folks, they’re just clicking away 100 miles an hour and if you move something all of a sudden they’re like, ‘What do I do?’” So, you get pushback when you try to iterate.
LEDGE: Be startup-ey.
JASON: Yes. We now consider ourselves enterprise software. We have longer release cycles. We communicate what those changes are going to be well in advance. Offer training, et cetera. It’s a much different process than you wake up one day and your Facebook feed is totally different and everyone complains but then they go back to normal. That doesn’t happen in our business. We would get obliterated.
LEDGE: That’s really interesting. Did you end up going down the direction at all of, you talk about taxonomies for the politicians and such for your original project. I wonder, have you come to any common data models or taxonomies around trying to integrate all the asset management systems if you have to communicate with different ones?
JASON: We have. What we’ve done there is, we’ve abstracted on our part what we call an integration layer. The primary problem that you’re solving there is a data mapping problem. So, we have a data model, they have a data model. How do you map variable to variable in a way that scales?
I think to date we’ve done I want to say like 70 integrations on the backend. A lot of those systems are the same. We do a lot of Cityworks integrations. We do a lot of Lucity integrations. But there’s also a lot of other smaller back-office systems like Maintenance Connection, for instance, or something like that. Every one of their data models are different.
What we do is, when we build an integration connector, we leverage a lot of the same code to do the mappings. Obviously, some of them are using web services, SOAP. Some are using REST. Some are using JSON. Some are using XML. We have a library of existing code that we can kind of copy from, at least from a template standpoint, and then once we have that integration up and running we can support future versions within the software.
But the primary problem that we came into was variable mapping, especially with systems that are highly configurable. You think of a sales force where you can add custom fields and data validation on X, Y and Z, you kind of learn to adapt to that. So, we’ve built out a fairly robust model that supports all that stuff on our backend so we can plug in.
We don’t want to write a sales force integration custom every single time, so we don’t. We’ve developed code to where we can kind of configure our own system to adapt to their system, essentially is what we’re doing.
LEDGE: Right. That’s another critical point that any founder should take seriously. Even when there’s data standards, there’s a lot of versions of data standards. We see this in healthcare all the time. It may be HL7, but which HL7? Particularly government likes to have different versions of everything.
Excellent. Great learnings and great stories. Put on your dreamer hat again. What’s next?
JASON: We’ve actually started positioning and developing IP around the notion of a single master application for your city. So, going beyond service requests – and we’ve already started to make strides in this area.
We don’t believe that a one-off, single use application is really all that interesting. Simply pull out your phone and report a service request. That’s great, it works, but for us I want to be able to do everything with my city on my app. So whether it’s I want to renew my dog license and I snap a photo of my dog, enter in its breed and color and all this stuff. Then when maybe my dog gets lost, I can send an alert and that goes out to my neighbors. It alerts the animal services department.
Things like disaster management. We’re in LA. We get earthquakes. You can do all sorts of applications that you really don’t think about until you need them, and really that’s the key. Is that, we see the phone as the communication channel between the city and you. So if there’s an amber alert, if there’s a road closure, if there’s a new farmers’ market opening up, you’re going to use your phone to interact with that.
There’s all sorts of stuff the city can do to incentivize you to do that. To download the app and to engage. They may engage local businesses if they’re trying to drive downtown business. They may say, download the app we’ll give you a free bus ride and a $5 voucher for the yoghurt shop, whatever. All through the app, one stop does it all.
In LA, last time I heard there were 26 different apps to do various things. They even had one app just for streetlight reports. They may have upgraded that since, but when we first looked at it every department had their app. There was no mobile digital strategy.
What we’d like to do in our big dream is to have a single app that facilitates every transaction the citizen has with local government. Parking, feeding meters, paying parking tickets. All sorts of stuff. Unlocking the tennis court. If you really want to get into the stuff, you use your phone with smart locks to pay for tennis court, unlock the court. Or if it’s like a bike or something, bike sharing apps. All that stuff.
LEDGE: I already want it. I’m having trouble with some water things. I would really like those people to pick the phone up, so I hope you get Fairview, Tennessee some day.
Jason, this is awesome. Great, great stories. I love hearing it and thanks so much for joining us. I know the audience is going to take a lot away from it.
JASON: Thanks for having me. I really enjoyed it.