How can supply and demand affect the pricing of a product? In this week’s podcast, we had a conversation with Vishal Kapoor, Senior Director of Product at Shipt. He delved into the intricacies of dynamic product pricing strategies, the appropriate timing for implementing this approach, and shared valuable insights gained from his experience.
Lily Smith (00:00):
Hey Randy. Bad news I’m afraid, there are so many podcast interviewers around that your pay is going down.
Randy Silver (00:07):
What? But the demand for content on product management is skyrocketing right now. People want new insights every week. It’s madness.
Lily Smith (00:18):
Okay. Okay, I suppose so. Let me just tweak the algorithm and I’ll give you a bit of a boost again.
Randy Silver (00:24):
Lily Smith (00:25):
And in case anyone didn’t catch on, this week, we are talking about pricing and specifically pricing that changes in response to the market.
Randy Silver (00:34):
Yeah, thank you. That makes more sense. So this week we’re chatting to Vishal Kapoor. He’s a Director of Product at Shipt, and he’s filling us in on how and why Shipt changed their pricing strategy. It’s a very dynamic chat. See what I did there?
Lily Smith (00:53):
The product experience is brought to you by Mind the Product. Every week on the podcast, we talk to the best product people from around the globe. Visit mindtheproduct.com to catch up on past episodes and discover more.
Randy Silver (01:04):
Browse for free, or become a Mind the Product member to unlock premium content, discounts to our conferences around the world, and training opportunities. Mind the Product also offers free ProductTank meetups in more than 200 cities. There’s probably one near you.
Lily Smith (01:23):
Hi Vishal. So nice to have you on the podcast with us today.
Vishal Kapoor (01:27):
I’m so excited to be here. Thanks for having me, Lily. And Randy, nice to meet you.
Lily Smith (01:31):
So today we’re going to be talking a little bit about pricing and a very particular type of pricing, but before we get into that, it’d be great if you could give us a real quick intro into who you are and how you got into product and what you’re doing today.
Vishal Kapoor (01:51):
Sure thing. So my name is Vishal. I am a Director of Product at a company called Shipt, S-H-I-P-T. This is a delivery marketplace similar to, you may have heard of Instacart or DoorDash or Uber Eats, so delivery marketplace similar to them, and it is the largest acquisition of target.com in their history. So this is a marketplace that operates in North America essentially, and we do deliveries. The use cases that we support are shopping and delivering and order for a grocery order, as well as a pickup and delivery order.
So for example, if you wanted a computer delivered or some electronic item delivered, we have partnerships with places like Best Buy and all these electronic providers, for example. Different electronic providers, best Buy, Home Depot, et cetera. We also have partnerships with chains like Costco and Safeway and Whole Foods and these other places where you can go and do deliver groceries.
The marketplace itself is divided into three sides. So there is a customer that places the order, there is the store where the order is, where the inventory is and the order is shopped from. And then there is the gig side. It is a gig based company like the others where the shoppers are going and fulfilling, actually doing the shopping and doing the fulfilment of the order. So they would go and shop at the store and then go and do the delivery. So I am on that side, on the gig side and a couple of big products that I own.
One of them is our equivalent of Uber pool or shared rights. So in the Uber world, if you were going from point A to point B, Uber could actually pull multiple drivers, different riders into the same car together. Similarly, we have a product called bundling where if a shopper is going from a store to deliver one order to a customer, we could give them multiple orders which need to be delivered around the same time, in the same vicinity, potentially at the same store, maybe at the stores nearby and so on. So we improve our unit economics through economies of scale. So that’s one.
And the other one which is more relevant to the topic that we’re going to talk about today, is I and my teams own all of the earnings or all of the pay that goes into the pockets of the shoppers who, when they finish a job for us, anything that they get paid after they decide to come and work for us is something that my teams own. Just to give you a sense of impact between the two products I talked about, it’s every year our P&L surface area is about a billion dollars. Between 2021 and 2022, we spent about a billion dollars to the shopper. So it’s a lot of power, a lot of responsibility to quote Spider-Man, I love that quote. So it’s a lot of science and art and a lot of thoughtfulness that goes into building and launching the things that we do, which is one of the reasons why I love talking about this topic. I appreciate the opportunity to do the podcast.
Very quickly before this, I was at Lyft, which is another, which is number two competitor to Uber. And I was also working on dynamic pricing, surge pricing type of problems at Lyft. Before that, I was a PM for a gaming company. We have a game here called Words with Friends. I’m not sure if you guys have heard of it, you play Scrabble with other people. It’s a mobile app, so I started my PM journey there. Before that I was an engineer and an engineering manager. Back in my older job, I used to write code for search engines like Bing is very popular nowadays. So Bing, Yahoo Search. And I started my journey with amazon.com. So that’s a little bit about me.
Lily Smith (05:37):
Thanks Vishal. That’s awesome. I’ve actually used Words with Friends, so that’s really cool.
Vishal Kapoor (05:43):
That’s awesome to hear.
Lily Smith (05:46):
So you’ve mentioned there about the dynamic pricing projects that you’ve worked on or products that you’ve worked on.
Vishal Kapoor (05:54):
Lily Smith (05:55):
Tell us a bit about what dynamic pricing is.
Vishal Kapoor (05:59):
Sure. So price variation is essentially a simple concept. If you think about it, changing prices has existed since the beginning of time and the beginning of money essentially. I think one way to simplify and simplistically think about the problem is think about somebody charging a premium for something that somebody wants urgently. So let’s say that you place an order on amazon.com, and you want the order instead of being delivered next week, you want to deliver it earlier than that, or you are shipping something to somebody else and then you want some more immediate delivery or instant delivery for example. So that happens. So in the cases where there is an urgency where somebody is looking for a premium service, the companies will charge you higher prices. So that is one form of dynamic pricing.
Another example that I like to give is your happy hours. When you go to bars, when you go to restaurants, there are times of the day, periods in the day where they want to discount so that they want to move their inventory, so they’re essentially lowering the price. So one example is where the prices are raised. One example is where the prices are lowered. When you do that, taking into account market dynamics, which is you’re taking into account things in real time. The true meaning of dynamic pricing is you’re actually aware of your supply, your demand, your inventory. These are the three variables usually in real time and you’re adjusting prices to balance everything in the marketplace. So that’s what dynamic pricing is, has existed since the beginning of time though.
Lily Smith (07:36):
And typically, is it something that’s displayed to the end user so they’re aware of the fact that this pricing is changing?
Vishal Kapoor (07:47):
Yes, it depends on, yes, the short answer is yes. But it also depends on who the user is. So for example, just to give you an idea of how it works in marketplaces or taking us back to foundations to the fundamental. If you think of how an economic system works, economics 101, how it works is imagine that you have a bunch of suppliers, people who are giving a product, and you have a lever as a company that you can change prices. So suppliers are probably going to come more and more back to you if you raise prices, right, because they’re going to make more money that’s obvious.
Imagine on the, now put your other hat on where you are a customer and you’re on the demand side. Now if the price is very high, you’re probably not in the game depending on what that item is. But as the price start to decrease, more and more people will come into the marketplace and they want to purchase at a lower and lower price, just generally speaking. So if you try to plot this, if you think of this and in your mind you make an XY plot, you’ll basically, and the price is your, let’s say x-axis, where it is increasing as you go out and then your units of supplier demand are increasing as you go up, let’s say that’s the y-axis, you will then realise if you plot the demand, you will realise that the demand basically starts very high when the price is zero and then it slopes down because as the price increases, demand decreases.
Supply is the other way around, which is supply starts at zero. When you’re not willing to pay anything, nobody’s going to really come and work for you or give you something. But as price increases, that line goes up. So it forms an X in your mind, if you imagine that. And the point where they meet the two meet is in a very simplistic scenario, in economics, it’s called the equilibrium point, which is when your supply and demand are really equal and are balanced, which means that that is a price point where people are willing to come and sell product into your marketplace or provide a service into your marketplace as well as on the other side, customers are willing to buy that product at that price point as well. So that’s called the equilibrium point. Anything above or below you are really in a disequilibrium. So there are situations where you have more supply than demand, meaning there are a lot of, and let’s say in the case of Uber for example, a lot of drivers, but there is no rides. So they’re just idling out. So that we call that as oversupply.
And then there is a converse situation where you have a lot of demand but there is nobody to come and serve it because there isn’t supply and we call that undersupply. And undersupply is one of the scenarios where, coming back to what you just mentioned, Lily, is a problem where it becomes very severe. Which is a problem where you start premium pricing, which is a problem where you start something called surge pricing. So surge pricing is a part of dynamic pricing, is a technique of dynamic pricing which uses market dynamics, the supply demand curves that I was talking about into account. But then you vary prices depending on how do you want to manage and restore the equilibrium back to the marketplace. So what happens typically just one more level deep is when you surge your demand, imagine yourself as a customer, the prices are very high.
So let’s say if all three of us were in looking for an Uber ride right now and the prices were very high, maybe I wasn’t willing to take it at that high price. Maybe I drop out, maybe Randy, he drops out, but maybe you are. So you are the one who gets that ride at a higher price. And on the other side, the way these platforms work, including Shipt by the way, you can translate that into our marketplace as well. On the other side, what happens is that higher prices are also shown, to answer your question, are directly shown to the people who are coming and doing the fulfilment, meaning the shoppers or the drivers. So as a shopper or a driver, I have a higher incentive because I’m making more money than usual to come and actually work versus on the demand side, it helps tamp down the demand a little bit because the prices are high. So it’s used as a lever on both sides, the prices are shown on both sides. That’s how it’s done.
Randy Silver (11:56):
You make it sound really simple to apply, as in you just feed in lots of data and calculate the equilibrium point and bang, it’s easy. And I’m guessing, I mean there’s going to be market conditions where it doesn’t make sense. You’ve got to have critical mass and things like that. So before we talk about your decision of implementing dynamic pricing at Shipt, just a more general question. When is it right to start thinking about putting dynamic pricing in? I mean, you’re not going to do it when you first get to market before you have any traction I’m guessing.
Vishal Kapoor (12:30):
Yeah, that’s a great point. Yes, you need a certain critical mass to be able to price dynamically. And I do think that pricing generally, in general, without talking about a specific strategy of pricing, pricing in general, and the way we think about this at the company as well is it’s usually a non-reversible door. If you update your pricing model, if you go from let’s say a one-time purchase price, like Microsoft used to do for example, sell Microsoft Word or Office Suite, which you installed on your computer. If you go from that to a cloud subscription, it’s a very different pricing model. Subscriptions are very different from an upfront payment for example. So these strategies are usually non-reversible. So you need to be thoughtful about how are you actually, are you at a marketplace or are you as a company at a point where you are even mature enough to do dynamic pricing?
So I think you’re exactly right in that how do you actually get there? So I think what happens is that there are a bunch of different strategies, and usually what happens is that the way it progresses is usually what happens is that companies will start with something called cost of manufacturing. So if you’re a small company, let’s paint this picture out. You’re a small company, you make some goods, you sell them, it does take some money for you to actually produce those goods into the marketplace to make those goods and sell them. So there is a cost of manufacturing, there is a cost of sales, there is a cost of marketing, and if you are a small company, all you want to do to do right now is breakeven. So you calculate all your costs and you say, I’m just going to slap a margin on top and I’m just going to sell my product at that price point. That’s fine.
The other way that people do sometimes is, so that’s almost how companies start. That’s almost the 101 of how companies start pricing. They assess their own cost and do a margin on top. The next level then is now let’s say you grew a little bit and now you feel like you are a competitor and you have other competitors in the marketplace. At that point you you’re going to start benchmarking yourselves versus what other people obtain. So that form is called competitive pricing. Meaning if Uber rides are at a certain point, for example, a certain price point, ignoring dynamic pricing. But if Uber rides were priced at a certain price point, maybe the Lyft rides have to be priced competitively to them because otherwise the demand will potentially, if Lyft is a lot more expensive than Uber, probably a lot of your market, a lot of your riders, a lot of your demand is going to shift into that competitor versus you. So at some point when you grow, when you’re big enough, you have to start benchmarking competition.
Then I think the next thing that comes up is something called value-based pricing, which is again, in economics the concept is called what is somebody’s willingness to pay or willingness to sell, which is at this point we are talking about personalising the prices. So Lily, your willingness to pay might be higher. We took that example before. When the prices are dynamic, your willingness to pay that price might be higher than my and Randy’s because of whatever reason. At that point you feel like you are willing to pay a higher price. So that personalization sometimes down to the user, sometimes maybe at a level of a market, at an entire market. For example San Francisco prices might be higher than, for example, Boise, Idaho prices in the US because it’s a smaller town. So that kind of competitive pricing, so that kind of value-based pricing, that’s the third level that happens.
And then finally, when all of these things are implemented, then you come to this real time market dynamics, what’s happening in the market right now, you have developed enough critical mass off demand of supply, of enough rides coming in, enough drivers, enough riders, and taking Uber as an example that you actually can start changing the prices in real time or you look to start changing the prices in real time. So you’re right, it’s a layered strategy. You usually start from a very simple strategy, which is what does it take for me to make this? I’m just going to break even or cost a little bit more. From there, you slowly advance into dynamic pricing research.
Randy Silver (16:51):
Now, are there times when you shouldn’t use this? I mean, I’m guessing monopoly or monopsony situations, it might be even illegal to use it at times.
Vishal Kapoor (17:01):
Is it illegal to use surge pricing? At the end of the day, the companies, I’m not sure if there are laws that prevent you from doing surge pricing. I’m not sure if there is, at least not to my knowledge and I’ve been in the gig industry for a while. So I will say that there isn’t a legal precedent or there isn’t a legal checks and balance for this. What does happen is that sometimes these algorithms can cause sticker shock. So one data point for example is it all sounds hunky-dory, right? You’re charging higher and all of that.
There was a data point from 2015, I saw a study from Uber where when the price is surged, only 10% of their demand was actually booking rides at that price. Actually less than 10% was booking at that price. So guess what? You are not going to be able to make up all of your revenue, 100% of your revenue through 10% of that demand. It’s not going to happen. If it’s a $100 ride, and let’s say if it is $100 ride and there are 100 people losing a $100 ride and only 10 of them are going to book it, you are not going to now… 100 into 100 is $10,000, you’re not going to be able to make those $10,000 from those 10 people only. You’re not going to do that. You’re not going to raise the cost of every ride to $1,000. You’re not going to surge price to that.
So there is a sticker shock, and I think the long-term impact on the marketplace is that it does create churn, it does create problems of people coming back. It does create a trust issue. So that’s why companies in their own interest, in their own experience, they actually have surge gaps or caps to how much higher pricing goes. On our side, sometimes there are also situations where you have reverse options where for example, the lowest bidder wins. In that case they have floors because they don’t want something to go down to zero and cost a race to the bottom. So this is a forward incrementing price, surge price goes up, but there can also be reverse prices for example. So things like that. So there aren’t any laws, it’s a free market, it’s a free economy. There aren’t any laws go trying to govern how the economy should behave. But I think the organics and the human behaviour, the way humans behave, that’s how it dictates the business and how it works. So that’s a big part of it.
Randy Silver (19:30):
Are you ready for Mind the Product San Francisco conference happening in June?
Lily Smith (19:37):
If you’ve been before, you’re probably feeling a bit like me, desperate for your MTP Con fix. And if you’re new to it, this is the product conference not to miss. If you’re a product person looking to advance your career, expand your network, get inspired and bring the best products to market, then this is for you.
Randy Silver (19:56):
So what can you expect? Well, MTP Con is known for their epic lineup of speakers, renowned product leaders with invaluable insights and tactics to share. They cover a range of exciting topics that will challenge and inspire you to step up as a product manager always with something tangible to take away into your own product practise.
Lily Smith (20:18):
And don’t let location hold you back. Even if you can’t join in person in San Francisco, you can still be part of the action with their convenient digital only option.
Randy Silver (20:30):
This event is a must attend for anyone seeking to elevate their product management game. Find out more and book your ticket at mindtheproduct.com/san-francisco.
Lily Smith (20:48):
And you make a really good point there, Vishal, around I guess working out the financial viability of switching to dynamic pricing. So just moving on to how you did this at Shipt and how you worked out what that dynamic pricing strategy looked like. Was there a period at the beginning of this where you had to try and work with the CFO or one of the finance team to model out what that dynamic pricing would look like and how it would impact then the revenue coming into the business?
Vishal Kapoor (21:28):
Yeah, I think the obvious answer is yes. We wouldn’t just roll out the pricing model. As I said, all of these are different pricing strategies. We wouldn’t just switch a pricing strategy because it really just drives your entire P&L and the surface area of a billion dollars, you just don’t do that without thinking really hard about it. But I will give you an example of where Shipt was before and after, and this was a little bit before my time, before I joined the company, just before I joined the company in fact. In early 2020 and I joined in March. So what Shipt was doing until that point was what we call the commission based model, which is not very different from the markup based pricing that I was talking about, the cost of manufacturing pricing.
But we are a little bit different because we are a services platform, it’s a gig company and we ourselves do not manufacture inventory. We don’t hold inventory, we move inventory because other people, we are an aggregator platform, we are a marketplace, we are not manufacturing anything. We are actually helping aggregate all the demand and all the supply into one place and helping move inventory essentially. So it’s a little bit trickier like that, but the way we worked, the way our cost of manufacturing equation worked before we moved to dynamic pricing was, the way it works at Shipt, just to give you guys a quick background is as a shopper, let’s take the supply side here for a second.
Let’s talk about pricing on the supply side. So let’s say if we three are shoppers and we come to Shipt or Instacart or DoorDash take a pick or Uber Eats or any one of them, what happens is that you see as a shopper or a driver, you will see a list of options that are available in your area for you to be able to finish. So in the Shipt tab, there is a separate shipped customer app. We call that member, so Shipt member app. And then there is a separate Shipt shopper app, which is for the gig side.
So on the Shipt shopper app, once you sign up, if you log into the app, you’ll see a list of different options that are available to you to be able to shop and deliver. And then for each of those options that you see, we call them offers, for every offer that you see, you also see a price that you will see. So it’s called upfront pricing, which is you’ll see a price of what you will get if you take that order, take that offer and you deliver that offer. So far so good, right?
The way we did it in the past was we had something called a commission model, which was we had, I think it was a minimum $5 pay for every order that you would deliver was minimum five bucks. And on top of that, I want to say it was 7.5% of the actual basket, actual cost of the order itself. So Lily, if you placed $100 order for example, to keep things simple, this would be $5 plus seven and a half dollars, 7.5% of hundred, which would be 12 and a half dollars. So if I was a shopper and you were a customer, and if I took your order I would get 12 and a half dollars and so on. So it was a revenue share model, like a rev share commission type of model, a markup type of model if you will.
The challenge that that happened because of that was it created these things called unicorn orders is what we used to call it. And what unicorn orders were that there were these orders which were smaller in size like a laptop for example, or a cell phone or a mobile phone. They were very small in size, took very little effort to shop, but they would give you a disproportionately high payout because the cost of that thing was very, very high. And so what happened was, because in our marketplace at Shipt, because you have a choice of all these different offers that a shopper has, they get to choose what they want to work on because they’re independent contractors and not employees. It comes back to that point is we give them more choice because we want them to maintain their independence and control their own work on the platform for example.
Because of that, what we saw was there was a lot of cherry-picking of these unicorn orders. So orders which where shoppers thought were was more work, there were more items in the order, they had to do a lot more work, maybe they had to drive a little bit further, for example, were just not getting picked up. And the ones which were less work and a bigger, it looked like bigger basket or more valuable, the average order value, the order value was higher, shoppers would chase those orders.
So what happened was that as a result of that, it started creating inequity in our platform, because as a marketplace, we don’t want… We want families with kids, for example, to have the same access to this service to be able to get what they want, to be able to get things delivered versus a millennial who’s trying to buy a cell phone for example so we wanted to create that equity of service on our platform. So that was our old model. And so that was one of the reasons, coming back to what you were saying, one of the reasons as to why we actually moved to dynamic pricing. I’m happy to go into how that worked with you.
Lily Smith (26:25):
Yeah, so that was the main trigger for moving to that dynamic pricing. Was there anything else that-
Vishal Kapoor (26:33):
Lily Smith (26:34):
Was there anything else that made you look at that route?
Vishal Kapoor (26:38):
It was mainly that we wanted to create, so the principle that we as a company, the stand that we took and the principle that we went after was that we wanted to pay for the effort that was expended in doing something. And these unicorn orders were disproportionately high pay. The model, the pricing model was disproportionately high pay for a very little amount of effort. That did not feel fair. That was one part.
And the second part, as I alluded to before, was it was creating this inequity on the platform where families, for example, with kids who wanted just a bag of baby diapers, nobody was picking up their order because the price was very low, but we wanted them to be served and we wanted our demand to be served equally, not just chase the cell phones and the laptops, but also serve everybody else as well.
Variations of this have been implemented in other platforms like Uber and Lyft for example. They tried to avoid showing you where for the longest time they did try to avoid showing people where the pickup was before the actual pickup happened. And the reason was, oh sorry, where the drop-off was before the actual drop-off happened. And the reason was because if it was an underserved area, they did not want to bias the driver from not taking somebody into that area before they picked somebody up because otherwise that area would get underserved, and the job of the aggregator of the marketplace just like us, was to do right by everybody everywhere essentially. So that was the driving factor.
Lily Smith (28:05):
It’s really nice to hear how setting those product principles has really helped guide and shape the strategy.
Vishal Kapoor (28:13):
Lily Smith (28:13):
So once you’d made that decision to switch to that different pricing model, what were your first steps?
Vishal Kapoor (28:21):
Yeah, so the way, let’s work backwards from where we are now and how did we build the dynamic pricing model. So we actually use machine learning now, it’s a machine learning based model. And the way it works is that, again, coming back to the example I was giving you guys before is when you are a shopper in the app, you see all these different orders that are available. These orders have different attributes like how many items are you going to shop, where is the store? What is the time of the day when the customer wants the delivery? That’s one big factor.
When you are getting groceries, for example, there might be cold store items like milk and ice creams and as soon as you get the delivery, you may want to put them away, put it away essentially. So time slot mattered with whether it was in peak time or off-peak time for example, that made a difference and so on. There were a bunch of different factors.
So what our machine learning algorithm does is it takes all of these different factors into account and what it tries to do is it tries to, now today it tries to create a time estimate. Estimate the time that it’ll take for somebody to fulfil that order. So that is what I meant before when I said we wanted to pay not for unicorn orders, we wanted to pay on the principle that we would pay on what we think is the time to shop an order. And that’s how we would, that would basically set a time, set a rate for different, what the pay rate would be for a certain geography. We call it metro, for a certain metro what we wanted to pay, and then go from there. So that’s what it is. We’ve moved from that static pricing model, that inflexible pricing model into first into machine learning.
And as machine learning, as we started growing. As two years later, two years in, we started finding out, I’ll skip over a little bit, but the way we started doing it is first we had one version of the machine learning algorithm and it was calculating these estimates. And then what we were doing was we were looking at whether our shoppers were taking the time that we estimated. The way it was estimating by the way, was looking at learning from shopper data and trying to replicate what it was seeing. A lot of machine learning, if you think about it, is basically pattern matching. Even ChatGPT, which is out in the news, has learned from all these of these different data sets. It can even programme today where it has learned from billions of lines of code today, it’s really just pattern matching and replicating what it has learned over time.
So we were doing essentially the same thing. We were learning from how long was it taking for shoppers to actually shop and then train our algorithms and come up with an estimate. And then what we saw was that there were bugs. Sometimes it would for very small orders initially, we saw that it was actually overestimating for very small orders, but it was underestimating for very large orders, meaning the time was smaller than what the actual people were taking to shop it and so on and so forth. So along the way we had to change the technology, add different features is what we call it in machine learning to the algorithm itself. And slowly, and it is a a long tail problem getting that, it’s called accuracy, which is when you have the data that you’re training with, how accurately or how closely are your predictions matching the data that you actually see. It’s called accuracy in machine learning.
So getting it to a very high level of accuracies is a long tail problem. You’re always chasing that. The holy grail, if you will. So that’s generally how it is. Now the prices are dynamic or the estimates are dynamic because they depend on all these different factors. In fact, I’ll say this one last thing, the same order from the same store at a different time on a Monday morning would see a different time estimate to the same order, for example, on a Friday evening. Because guess what? There’s more foot traffic, there’s more people shopping on a Friday evening. And so when we look at real world data for exactly the same order, we would actually see that our shoppers in the store took longer to shop, therefore the estimate for that time should actually be longer. So that’s how it works, if that makes sense.
Lily Smith (32:35):
Yeah, that does make sense. It sounds really, really fascinating. How do you go about validating or you mentioned the accuracy there, how do you test to make sure that you’re accurate or are you getting feedback from your shoppers as they go to say, “Yeah, this was a good estimate,” or no it wasn’t?
Vishal Kapoor (32:57):
Yeah, so that’s a good question. It’s difficult for shoppers to say whether something is a good estimate or not. To be candid. This is one of those things, right? I’ll use the Uber example we were using before, how can we see? If there was a surge price on the rider’s side, on the demand side, how would any of us know whether this was a good price or not? For example, in the example we used, I did not book, Randy did not book, you booked, but how do we know whether this was a good price or not? So instead of asking people, I think when you ask, this is the thing with surveys about pricing and about pay generally, and this is a running joke that we have internally within the company.
We don’t actually run surveys for pricing or pay explicit surveys because every other survey that we run, the number one pain point that we get is you don’t pay high enough. That’s just how it is. By default, our number one pain point is that you don’t pay high enough. And that’s not news because that kind of sentiment is echoed around on social media. We have public forums and where our shoppers go and chat and we have Facebook groups and things like that. And that is a common complaint that happens. And in full candidness, right? Times are hard and inflation is at a all time high and there is a lot of things that government is trying to do to control it as well. So it’s not unfounded. The sentiment is not unfounded, but we don’t do it.
So what we do is we try not to ask people about what they say. We actually try to see what people do. So what we do is we try to see if we change a model. We do this through experiments basically. If we are experimenting with a model in a marketplace, a new model, something changes, how does our audience react to it? I think I gave you the Uber example, only 10% of the people, less than 10% book. That’s just not something that the company wants to do, right? Because at that point the company is actually not running a profitable business. They understand that.
But the reason why they do it is because they want to maintain a high level of service quality for the people who want to continue to be in the marketplace as well as get some more drivers into the marketplace as well. So it’s the same idea. We look at how the market reacts, we look at are shoppers still coming back to us, is our retention high? Is their engagement high? Are they continuing to shop the same number of orders despite the price change and things like that? So that’s how it works.
Randy Silver (35:25):
Vishal, this has been great. We are running out of time now. So I’m going to ask one more question and sorry, I’m going to put you on the spot for this one. So combination of two things, things, it may be the same thing, but for somebody new who’s trying, or somebody who’s trying to implement dynamic pricing for the first time, is there any advice that you’d give them and is that the same as the biggest learnings that you’ve gotten from this? Or is there something about getting started and then something for later?
Vishal Kapoor (35:54):
That’s a great question. I’ll give you a quick anecdote from one of the talks that I gave on dynamic pricing last year, around this time last year. And somebody came, I was talking on this subject of how to build dynamic pricing at a company. And after the conversation, somebody came to me and they said that during COVID, so a couple years out. During COVID, they had a marketplace where they had dynamic pricing, their surge pricing, but they never quite increased the pricing to, they basically had a price cap. They never went beyond that price cap to a point where they actually burned a lot of their investor money and their company had to shut down. And they were asking me, “Should we have done this? Should we have not done this?” After the fact, if you will. And my response to them was that you should have absolutely just gone for broke, right? You should have not had a cap. You should have just tried as high as the price would’ve taken you for two reasons.
Number one, you cannot predict what your services and how valuable that is. You should let the market predict for you. Don’t try to read the market’s mind, you can’t do that. Let them do it for themselves. And number two, if Uber and Lyft can do it and their surge prices, for example in COVID were egregiously high, now they’re starting to come back down. But it’s a problem that surfacing in the media as well. If they can do it, why wouldn’t you? Why would your marketplace not be able to survive high pricing? So I would say the three or four strategies that I mentioned before, start with markups, start with some of those strategies, competitive pricing, value-based pricing, and then eventually as your technology scales you get to it. But second, I would say do not be defensive.
And one slightly off-topic example that I give to people is, which I find extremely inspirational, is Airbnb was able to put people into other people’s homes, right? If you think about that idea in and of itself, it’s absolutely crazy. So I think you should have big, bold, ambitious ideas which you think will change the world because if you try them, you’ll be surprised that sometimes did. That’s what I’d say.
Randy Silver (37:57):
Oh, that’s fantastic. We really appreciate that. It’s a great story and it’s great inspiration. Thank you so much for joining us today.
Vishal Kapoor (38:04):
Sounds great. This was a pleasure. If anybody wants to keep in touch with me, please find me on LinkedIn. My handle is linkedin.com/in/vee. That’s V for Victoria, E-E, very simple. Love to connect and love to answer any questions.
Randy Silver (38:21):
It’s V for Vishal.
Vishal Kapoor (38:22):
It’s V for Vishal. V-E-E. So it was a pleasure being here. Thank you guys for having me.
Randy Silver (38:27):
And we’ll have that link in the show notes as well. Thanks again.
Vishal Kapoor (38:30):
Thank you so much.
Lily Smith (38:31):
The Product Experience is the first.
Randy Silver (38:45):
And the best.
Lily Smith (38:46):
Podcast from Mind the Product. Our hosts are me, Lily Smith.
Randy Silver (38:51):
And me, Randy Silver.
Lily Smith (38:53):
Louron Pratt is our producer, and Luke Smith is our editor.
Randy Silver (38:57):
Our theme music is from Hamburg-based band Pau, that’s P-A-U. Thanks to Arne Kittler who curates both ProductTank and MTP Engage in Hamburg and who also plays bass in the band for letting us use their music. You can connect with your local product community via ProductTank, regular free meetups in over 200 cities worldwide.
Lily Smith (39:18):
If there’s not one near you, maybe you should think about starting one. To find out more, go to mindtheproduct.com/producttank.