Scheduled for: Interviews/ Category:
Surge pricing may soon be something that we see for more than just an Uber ride, but what’s the real purpose of companies surging prices? And is it always bad for the consumer?
Alexander Shartsis is cofounder and CEO of Perfect Price, an artificial intelligence (AI) company empowering companies to make better decisions about pricing, profitability, and utilization. He is an entrepreneur and executive with background in business development, sales, product and engineering.
Previously, he was Head of Business Development and 5th employee at Drawbridge, brought on just after seed funding from Kleiner Perkins and Sequoia Capital. There he grew the company from pre-product and pre-revenue to a $10mm run rate and a $100 million valuation series B in 13 months. He closed both the first 12 advertising clients (e.g., Disney) and first publisher deals including MoPub’s first RTB exchange partnership, and helped build a world class business team.
Prior to Drawbridge, Mr. Shartsis was Director of Business Development at TripIt, from its Series A through its $120 million acquisition by Concur (subsequently acquired by SAP). There he launched the advertising business, lead the public launch of TripIt’s API, sourced, led and closed major distribution deals with Fortune 100 companies including American Express and Expedia, and created and managed relationships with partners including Apple, Google, Microsoft, and LinkedIn.
He advises or invests in a very small number of startups through his network. He advised TubeMogul (NASDAQ:TUBE, acquired by Adobe) prior to its Series A, and played a critical role in eBay’s acquisition of another startup.
Kim: Awesome, so good afternoon and thank you for joining us on dojo live today Wednesday the first of May 2019 happy Labor Day to the Mexicans out there, well what we’re about to talk today is about the case for surge pricing but before you do that let me introduce myself my name is Kim Lantis I am joining you from one of Nearsoft offices in Hermosillo,Sonora today I have with me my coworker Tullio Siragusa.
Tullio: Hey everyone, I’m here in Los Angeles sunny Los Angeles.
Kim: And Carlos Ponce who’s kind of running the back end for us and in Mexico City hey Carlos.
Carlos: Hi Kim, hi Tullio.
Kim: Yes, and today’s guest of honor so like I said we’ll be discussing pricing profitability utilization and how all that’s linked in with artificial intelligence AI here to tell us more is Alex Shartsis, I think I got that right.
Alex: You got it right.
Kim: Excellent, so Alex is the co-founder and CEO of perfect price welcome to Dojolive.
Alex: Thanks for having, excited to be here.
Kim: Awesome, so to get us started could you please tell us a bit more about you Alex yourself and your background.
Alex: Sure, so my background from California made it back to California after about a decade away I have a degree in an MBA from UCLA I have a undergrad degree from Dartmouth and a degree from the long school of economics in between and I’ve been in tech for about 10 or 15 years now turns out I’m gonna kind of old and yeah and so this is the second company I’ve started we do dynamic pricing using AI so really its value.
Tullio: If you’re getting old after 15 years I’m becoming ancient.
Carlos: I think I want to remain silent on that.
Kim: So now that we all established how young we are, why don’t you tell us a bit about more about perfect price so what exactly is your products and what problems are you solving what’s the need out there?
Alex: So, good question so we solve a really simple problem which is setting prices for large organizations so I think prices something that people take for granted and it’s actually turns out to be really complicated so so that’s that’s where we fit in so we help companies leverage our data and make really good decisions around how they price their products and services.
Kim: That is really interesting like if you don’t mind before we get more into the specifics of surge pricing what is some of the data points there I have to be taken into consideration.
Alex: So it’s interesting there’s a ton of data that you can take into consideration it really depends on the business so we work with companies to do all sorts of things everything from car rentals to automotive repair to airlines hotels I mean it’s sort of across the map even gas stations and so every business is different data that it needs to take into account but what we find is that the most basic data like one of your customers are actually doing it frequently just gets left on the floor like a lot of people don’t they you know they have an idea think about if you worked at a company for a long time you would kind of know that people are more likely to buy from you on a Monday or they spend a little bit more on Saturdays right but the dumb might not actually be reflected in how the company does business and that’s that’s what our software does it enables people to take advantage of that data if they already have a lot of the times and make better decisions.
Kim: Nice, I imagine it also helps reveal things that they probably hadn’t even thought of before and then take advantage of those ones.
Alex: Yes, yeah, both people are both surprised that the AI figured something out that it took them five years to notice and and so I also also surprised by things that I figured out that they never notice.
Tullio: I have a question, just I don’t know if it’s jumping the gun here but so what is the so I this is great insight that I can gain right but do I have the ability to set the intention on how I want to use it in other words is is it to improve customer successes that they get more margin what is what are some of the uses for this capability that you’re seeing.
Alex: Yeah, so I think every company different objectives right you have two companies in the same industry one one CEO will want to go for margin and the other will go for market share right so it really depends on what that company wants to use a software to achieve and then yes and we set up the software configure the software to do what they’re trying to do so I mean if you take the title of the session today about surge pricing really that’s obviously a word that uber made popular but in an uber or childless clinic’s goal with a surge pricing was to always have a car available like so their pricing strategy was there will always be a car may cost $400 but you can get a ride wherever you want to go one of you won’t go there if uber service is your area and so for them it wasn’t so they also do dynamic pricing and make it cheap the rest of the time to point people in to the platform which we can talk about later but but you know that’s one set of goals you have other goals you know a lot of the Clorox and candies of the world their pricing strategy tends to be much more focused on market share right they just they want everybody to buy their detergent or their their toothpaste or whatever it is that they’re selling and so their pricing strategies trying to figure out how do we how do we get this out and distribute it everywhere how do we priced it so that people buy ours and not somebody else’s even if it’s maybe not profitable on the austell so really it really remember the yes that software will you know we sit down people that helped them configure that or configure it for them so that they can drive the results that they need to see in their business.
Kim: But this perfect price have this ability to do the chicken versus the egg I think that was more your question as well like okay here’s the chicken what’s what are we gonna go for we’re gonna go for mark you know market share but can it also reveal I don’t know the egg of like hey you should be going for market share even though you think you should be going for fill in the blank like is there the reverse?
Alex: Yeah, I think we try to be nuanced about it I mean you guys you guys work with companies that you know that want software development resources right and sometimes they probably have not the best approach or not the best idea and so you want to deliver what they are asking for but steer them towards perhaps something that’s better for their business so there are a couple places where this comes up a lot one of them is costs so a lot of companies wanted take control of advising that haven’t done their homework on cost or have conflicting views of what should be included when calculating a price and so yes we’ll work with them there to help them make better decisions there are other times where they’re really set on market share and that’s what their board wants without with their executive team wants and so the best thing we can do is support them from market share and specifically our software will let them simulate other outcomes so he can show you this is what happens when your volume stays constant this is what happens to your profits when you add percent to your volume this is what happens your profits when you lose 5% of your volume but you’re a lot more profitable and companies can then make a decision about it and they use the software to make that decision okay but we if we’re okay giving that 2% market share we could we could make a lot more money do we want to do that and our software helps them make that decision.
Tullio: So Alex, is the platform do like simulation that the end users can look at before you actually apply the decision so is it it’s doing it on the fly based on the parameters you set.
Alex: So just both I think both are very key use cases so I mean let’s say you run a global car rental company right and I show up on your doorstep and you like ah this this AI pricing thing that’s I’m gonna be a good idea we should we should give this a shot right let’s see you have a ten billion dollar company you’re not gonna be I got just price it all like just knock yourself out they’re all theirs won’t mind if this doesn’t work right I mean part of dealing with pricing is recognizing the price is a really big lever to your growth and your profitability and everything else and so if it’s a big lever one should be careful with it it’s sort of the big red button that nobody wants to touch because it’s potentially really scary you know I mean it’s funny we talked to you one of the biggest hotel companies in the world a couple months back and you know one of their one of those senior executives who run strategy was paranoid about the pricing project because she was like this is a big deal like we know how important it is so simulations are a key part of any pricing solution you know that’s going to be effective and so we work with people to run through those simulations make sure they’re setting them up right making sure they’re getting to the people that they need to didn’t even see them because even even then right if you if you go and say okay we’ll make you know we’re gonna make another twenty million dollars next quarter by changing but by doing dynamic pricing doing surge pricing whatever that is and then I write and then and then your operations guy gets me is like we can’t triple our you know we can triple our capacity you know that that’s a way to do like you can’t do that right I mean so you need to you need to actually think it through and simulations are really powerful way of doing that I think what’s interesting about our approaches that’s the same system that then sets your prices so you’re able to actually see like Oh on the to euphonic our own company Oh on a Tuesday this is what’s going to happen in JFK or in New York City or what have you how do I feel about those prices I don’t feel about the results were supposed to generate is it realistic you know could my operational people actually do this right and then oh hey check it out I’m gonna make twenty percent more money I feel pretty good about that part so you always you always want to stimulate but but then you know once you’re running you don’t want to be stopping to think about it every Monday you know most a lot of our customers change hundreds of thousands if not millions of prices a day you you can’t look at everything under debt.
Tullio: Go ahead Kim.
Kim: Just saying that’s where they I guess the AI and the scalability comes in.
Tullio: But I’m curious though like I remember for example when SAPfirst came out many months ago, like 70% of those implementations failed right because the the promise was they just put the software in place and it’ll work but what they realized was they needed to customize and work with subject matter experts who actually understood the business requirements before applying some of the technology and software that’s obviously changed and evolved and they’re one of the best software company in the world today but my question is in order for you to effectively do what you’re saying whether you’re setting up simulations for different verticals and marketplaces are you also providing subject matter experts how are you going about that hurdle understanding what’s needed.
Kim: See, you just mentioned dynamic pricing which I understand to be synonymous with surge pricing which you had mentioned, oh I froze, I’m sorry, did I missed something? I’m back.
Alex: You said, dynamic pricing and surge pricing and then you froze.
Kim: So you just mentioned dynamic pricing and previously you had mentioned surge pricing that’s my understanding mites my understanding those two things are some synonymous.
Alex: Surge pricing is a form of dynamic pricing.
For financial dummies like myself, could you put surge pricing into layman’s terms for us?
Alex: Sure, so surge pricing is essentially raising the price when there’s more demand I think hence surging the price I think it’s kind of gone out of fashion as a term but it’s a fun title for a front you know webcast because because it because it has gone out of fashion for very good reason but I think I think dynamic placing more broadly is changing pricing to reflect the may handle or anything else right you can change pricing that’s because you are following your competitors and they changed price but that’s a dynamic pricing is just prices that change might the example I always go to is jet GE sells a lot of jet engines or Rolls Royces it’s a lot of jet engines those are not to endemic prices you know they sell here 100-year Boeing has a price list that they get every January that says you know this jet engine costs X million dollars and it doesn’t change in March doesn’t change on a Monday or Tuesday right that’s never going to be a dynamic pricing problem but the airline seat the seat on the plane on them United but that’s always dynamically price now is it a search price like you could say that you want to fly the Hawaii over Christmas that’s surge pricing but you know they didn’t call it surge pricing they just call it the price.
Tullio: I can search pricing like it reminds me of uber right it’s like it’s something that’s instantaneous and it could drop off five minutes later I guess, before I ask my question I’d like to meet me back up a little bit cuz this helps myself helps the audience understand that sort of Genesis behind this idea what was that moment it’s like okay I wish we had this what was that you were the company was trying to solve in the beginning that you saw a need in the market places I think that also help better understand this.
Alex: Yeah, so I think it’s the way we approach it is a little bit of a leap so you have to know before I started in this company my co-founder and I were at a company called drawbridge so is one of the three or four companies that kleiner perkins at Sequoia Capital seed funded and they’re doing quite well it was you know I was there the fifth employee and and so we were using AI and machine learning to buy and sell ads and but mostly I guess the buying of ads and it would sell them under normal old school contracts but but what what we saw was that they were there are these companies that existed then old school media companies and new ones like rocket fuel which at the time is public that had people and so these people would I mean I rocky the CEO rocket fuel showed me the soccer field where everybody played you know intramural sports after after work or even maybe during the work day you know they had hundreds of people to optimize their ads we had one guy my co-founder and now in AI a machine learning and we crushed them I mean it wasn’t even close and so when I looked at how we did that and then a couple friends of mine were running different pricing organizations and bigger companies and they were doing it with Excel and humans the way that those legacy media companies would I mean I was like this is this is just wrong you know I mean we have this great technology why aren’t using that and one of those friends was not here’s a chat but you should go you should go do that like we the world needs this needs this and I need it like oh go build it right now and so so that was really where the insight came from was that you just you have a list data as a mate’s as a normal company the normal enterprise in these other areas you can use that data to make really good decisions and pricing was not one of those places where it was existed to make good decisions.
Tullio: Was that Randy you talking about that rocket fuel?
Alex: It wasn’t Randy actually, I mean I like, I like the rocket fuel does it was just it was interesting it was it was and he was like speaking another language you know, we really want you know and also we were we were focused on mobile and back then they’re like oh yeah we have this guy he doesn’t mobile like part time because in a relative thing for us you know it’s like a couple cons asked about and you know like it’s not going to be doing but like we’ve put a guy on it because we felt like we had to write and obviously like you know we’ve seen that movie play out well.
Tullio: But it sounds like you kind of have created similar to the real-time bidding platform in the at tech space for the business for the reason that.
Alex: That’s exactly we we’re a participant on the like rule on the first mobile real-time bidding buyers and so we took very similar algorithms technology and obviously don’t have you know you buy an ad right that’s the consumer doesn’t care nobody complains about surge pricing for this click you know that’s they do but you know there’s nobody to complain to oh that Google doesn’t pick up the phone but but you know in in real life it’s a lot more complicated what we’ve learned is it that the technology is just as capable and it’s having that sort of human-computer interface so that so that people are comfortable with what the algorithms are going to do at Tec is kind of well unless it’s like the algorithm to do what they do and we clean up the mess the next morning but in you know if it’s if it’s your pricing and your company actually about it was quite a bit you know you’ll have you’ll have upset customers are upset staff and so so we’ve added a lot more on top of the technology usually.
Kim: That’s actually was going to be like one of my questions I was gonna save it for later but I think it’s a really nice transition earlier you mentioned I mean companies that have 40 billion dollar-a-year profits or to me that’s like a number that I cannot even physically unable to even comprehend such numbers um and then you transition this nicely into this idea of artificial intelligence and cleaning up the mess later and like so what is it like what’s your process to get people when we’re talking about such large states to buy into to trust a machine like artificial intelligence how do you accomplish that in your company?
Alex: That’s a great question, I mean I think you know I think the expression is poco a poco good start with the whole forty billion dollars or you start let’s start with your run you have 10,000 gas stations let’s start with three right let’s go with one and make sure the prices actually show up on the pump and then we’ll then we’ll do three and we’ll see how they do and if that works we’ll do five so I mean I think I think there’s there’s a mindset shift that has to happen and that’s the hard part you know you have to most companies the gossip destinations are funny because I think the people who manage place a lot of them are paranoid about like the CEO is gonna do if he drives by their station and they don’t they don’t have the quote unquote right price look at a phone call sitting your desk and like the CEO of Shell will call out back why is why are we charging $1.99 here you know it should be I think it should be $2.05 right so I mean there’s you know people everybody’s opinions I think you have to move you know Jeff Bezos has never walked by somebody’s desk and said why is this book 999 right I mean like there’s there’s a respect for like data at a company like Amazon and a lot of a lot of companies just haven’t made that leap yet and so what our technology does is it you we help you make that leap by making it easy and so I think once you’ve said okay we’re gonna turn some of this over today and see how it [INAUDIBLE] once you can’t make that leap we’re not working with if you can have that kind of trust in the system then there’s a very very straightforward that well-tested improve process for proving that it works so that you’re not you’re never like let’s just hold my beer let’s let’s do this without forty billion you say you can make us ten percent more profit like show me you know I mean let’s show you but I’m three you know in three stations or three locations or a small line of business right let’s not you know let’s not do something it doesn’t have the ROI of the long term but like let’s start small prove it out and then so.
Tullio: So I have a question around that because early in in the machine learning space many companies came out with personalization specifically in the mark tech space for example but it required large sets of data for it to work like in massive sets of data for it to work and and it became a little bit of a challenge to try to do small set tests or begin small has that been solved since then is that what you’re saying is that you know yeah we can start with three stations for example instead of having a test 10,000 how are you managing through that you do you look at the entire sets of data and then you kind of parse it down or you have another methodology that the algorithm that you’ve adopted.
Alex: Yeah, so the two aspects to that one is the technology and being accurate on small data sets so my co-founder did a lot of his work on the Microsoft search bar on Twitter and I know we’re on Twitter now people don’t click on Twitter you just you don’t click and that’s how you train search algorithms is the clicks so no clicks and things are very difficult so our tech knowledge was built for very low data volumes so like it’s the same thing your toolbar I mind yourself like it takes the Microsoft search result and then it modifies it based on your browsing history and the stuff he knows about you and so that’s always a much thinner data set then how many people clicked on the Apple site or something like that right what have you done on the Apple site well a couple things right you it’s it’s guessing but it does that in an intelligent way so our software leverages those kind of technologies and we patents on how we take this sort of really thin data set and turn that into a very accurate stable model or set of models so that’s one aspect of it is the gifts that I know and not and that is not that was very hard to home that I’m thankful for the team area up here their ability to solve it I think I think the other side of it is you can you can start with the pricing on three stations but have trained it on five or 10,000 right you don’t you don’t mean that where you implement it does not need to be the same as what you’re analyzing and so what we find is that most of our customers at first step is just they do the simulations and they wanted to sort of touch and feel the data they want to like the what the prices would be they want to put those they have my pricing team in the president’s you can so they can have opinions about it and usually what happens is they complain they’re like oh this Memorial Day pricing just looks a little bit it’s like okay what about everything else like the rest of it’s fine it’s just really twos this Tuesday’s problem and everything you own 98 percent of its fun and that’s and that gets uncomfortable with it and then you say ok not even though you’re comfortable with it everywhere we’re gonna focus on these three if you know locations and then 10 and then 15 you know and roll it out globally a big difference to the technology is that the old way of doing it was varied batch process so you would do the analysis on the three locations and then you have to redo it on the next three four is with AI once you get it right you know after we do it you did it once it works everywhere or almost everywhere and so you could hold on much more quickly but but so we separate that the data analysis part of it from the testing going live part of it because they don’t have to be the same.
Tullio: Hey, quick quick interruption anybody watching want to submit a question on Twitter it’s @DojoLive or if you’re watching on LinkedIn please submit a question Carlos checking and and monitoring and Carlos if we do have some audience questions please jump in and dropped us and let’s let’s get those answered speaking of what you just mentioned so are you doing like dynamic trees in terms of how your decisioning are so it’s continuously adapting and changing based on outcomes or is it done you know there’s a study it’s done it’s implemented and then there’s a change that looked at before accepted is that if someone says I want continuous dynamic pricing search pricing is it automatic or they’re still menu intervention.
Alex: Yeah, so the it depends the customer both are possible so if you’re we work with companies that sell cars right like an actual you know car you go to dealership and buy and and those those companies just you don’t want to change the price of a car three times a day but while you’re walking around a lot suddenly the car got more or less expensive like thought it would probably just be weird might get you to buy it but it would probably just be weird right whereas a rental car price will change every 15 minutes you know, Amazon change their prices every 10 minutes or five minutes right so so it depends on the context if you’re changing price every 10 minutes you cannot possibly look at that price right if it’s any significant number of things your pricing or as if you changed price once a week you can put a human enough in that process and it doesn’t doesn’t really slow you down eventually you probably decide not to but if that makes a customer comfortable fun you know how could you do it and push the button so it really it depends on the on the use case if we take the title of surge pricing like you know you can open the uber or lyft app and you can see the price change in front of you a human is not reviewing every time right, a lot of different pricing is left right you know I mean that’s that’s algorithms it has to be our general approach is whatever that time period is or we eat today and depends on the data an hour or 15 minutes you we can reach rate the model can retrain itself and learn from the new data which is really effective in fast-moving industries like the Ubr use case that’s where you know a lot changes in five minutes and you know drivers come on the platform drivers go off the platform cars move around passengers come on and off the platform and customers changes it makes sense to have the new okay you have the AI have a new opinion of the world as it is you know are you going to how many cars is Chevy gonna sell this five minutes and you know West LA not like not enough to change what it thinks the price of the Chevy Tahoe should be but maybe this week it will write maybe three days or now it will have a different opinion.
Tullio: So for potential clients watching this who are the right or the most perfect vertical markets that you want to serve.
Alex: You guys, who work so we focus a lot on transportation I guess there are two dimensions one is you know you have to have a large business because you need a lot of data and you need to be able to invest in in done in dynamic pricing so if conversely if you have a small business you only one gas station you to stand on that corner and drive around and look what your competitors are doing and do just fine like AI is gonna do better than that but the ROI on a gas station is just not one gas which is not there you know to make the investment in integration technology the rest of it so for us it’s really large companies with complex data generally speaking in the transportation space or an adjacent space where there’s you know like hotels airlines things things that were you have high fixed costs you know and lots of data right lots of transactions it’s funny I had this great conversation with the guy who does pricing and the CEO of ring the doorbell company two years ago they’re coming out of the new version and and you know and it was yeah I mean have people calling me just through my network to ask pricing questions and it was one I was like but you put it in a Home Depot man like you get there’s no there’s no dynamic pricing when you deal with Home Depot you got one price so you know you got it I mean there’s strategic decisions to make about that price right if you have one product you don’t need software yeah not that complicated.
Tullio: Right, okay so we’re kind of coming up on time and it’s always this always happens when we are having fun.
Kim: It’s not kind of, we have.
Tullio: I used to be you could do an hour-long interview but now if you haven’t noticed even at conferences workshop nothing lasts more than 30 minutes because the attention span isn’t there so let’s enjoy 30 minutes because I’m sure that’s going to get down to 15 in the next few years but anyway speaking of timing and lessons learned what are some of the things that you’d like to shoot here anyone who’s sort of contemplating this journey of doing a start-up and building a business words of wisdom that you’d like to share with them things that you’ve learned that through this journey we’d love to hear them.
Alex: Sure, I mean I guess with wisdom a couple things that I haven’t done this several times as an employee and as a founder to me the most important thing has worked with great people both great people from a professional standpoint but also great people that are really good at what you know I both really go to what they do and they’re people that you enjoy spending time with and then have things in common I think for us at perfect price are we really focus on finding people that are doing what they do and also really like learning it’s you know we’re an AI company it’s all about machine learning but it’s also about people learning and it’s fun you get to learn about new industries and companies new businesses new ways of doing things and then always raise more money than you think you need.
Tullio: Right, well I appreciate you taking the time with us it’s been fun love to continue it stay in touch and see how things progress for you Kim if you want to close us out.
Kim: Yeah, I think you you did really well thank you for your words of wisdom once again everyone this is Alex Shartsis co-founder and CEO of perfect price thank you for time today Alex.
Carlos: Before we go, but before we go, let’s not forget what we have for next real quick and for that is we are going to be speaking with Utpal Kaul the new product incubation from new product incubation of carlson wagonlit so that’s that’s trouble tech waiting gentlemen so keep an eye for that Wednesday 1 p.m. Pacific.
Tullio: Anyone is in the startup space that’s thinking about a startup in the travel tech space they should tune into that show I’ve met Utpal in Miami, they’re doing some interesting things and I think we’re going to have our head of R&D joining us on that cause well in that interview as well so I’m excited for that interview next week.
Kim: Oh good, well that’s it for me thank you guys thank you for your time today Alex and have a great rest of the day everyone.