1% of revenue for any company earning in excess of $10k per month seems excessive. But I guess if those clients are also seeing substantial percentage increases in revenue it would warrant the 'data fee'. This is classic supply and demand on-the-go pricing, but from my initial glance the site doesn't really explain where they have tested it, and what real results were measured.
It is actually 1% of revenue for all companies: the lower-priced brackets have caps that make it work out as 1% (or higher, due to the granularity) across all brackets.
My problem with this is that there is no way that this algorithm is really that impressive to warrant what might be a substantial amount of someone's margin: if you are finding yourself spending a ton of money on this algorithm, you can easily re-implement it... based on both the video and the FAQ it sounds like they are doing simple hill-climbing, which should take a few hours to implement (and which, I will however point out, isn't even anywhere near optimal for this kind of problem).
It's not hill climbing. Gradient descent has some severe limitations.
Also, I have mentioned it before but if you are charged 1% and someone makes you 2%, I don't understand why you wouldn't buy from them. (On a side note: we earn our sellers more than 2% on average)
Rereading what I said, I am not certain of the confusion, but I will expand: you can't evalulate the price of your product using the gross value it provides in isolation, you have to evalulate the price of your product using the net value it provides in comparison to competitors.
If I ended up in the situation where I was paying you hundreds, certainly if I was paying you thousands, of dollars a month, I find it impossibly difficult to believe that I couldn't get more "bang for the buck" by spending the time to reimplement the features I am getting from you.
(As an entirely unrelated argument, the optimal margin might not even be 1%: to charge a percentage of revenue for something that doesn't have any scaling terms in proportion to the dollars being moved is, to put it lightly, "bold"; my business, for example, probably couldn't afford 1%, yet I routinely move tens of millions of dollars... and I certainly, at the price of hundreds of thousands of dollars a year, would have no trouble reimplementing your setup, and in fact have done similar things as side projects on various occasions.)
Customers that do tens of millions of dollars a year tend to call us and get a lower price from us. They don't use the signup box. It's common for enterprise companies to self select by calling for special pricing and in most cases we oblige. We want to make a fair deal for all parties.
On the other hand, I get it - You would never use us, you would try to build it yourself. I can respect that. I get like that for a lot of stuff I see.
The founding team's past experience is in algorithmic asset valuation, commodity trading and machine learning. We are life long learners and are dedicating our time to find better ways of helping ecommerce stores pick better prices. We have several years of experience in workflow automation and we apply that to our current business to make our product as cheap as humanly possible and pass those prices along. Our closest neighbor in the pricing game charges hundreds of thousands of dollars in fees because they use so much human labor to do things.
Saurik, if you ever want to do one less thing and want us to do your price testing. Please give us call, I would love to find a price that makes it worth your time.
You seem to have missed the point of me providing that number (something I take some of the blame for, as I was explicitly mixing two arguments; I did go out of my way to attempt to differentiate them, however): I was demonstrating that even at an incredibly large extreme with regards to absolute revenue, the percentage margin doesn't really become affected; when I was moving tens of thousands (as opposed to millions) of dollars, the answer would have been the same: you can double my revenue, but at 1% you are talking about most of the magin that I'm currently playing with for my not-entirely-uncommon business model (where my upstream obligations, fees, and taxes, scale nearly linearly within my plausible target price range, with respect to the ticket price of the item being sold).
It isn't because I would be an "enterprise company" that I would need special pricing, it is because your pricing model makes a fundamental assumption about the kind of business model that the other person might be engaged in: if you had a "percentage of profit" model--even though you would still be in the "this is too expensive: I could hire a full time developer and a CS grad student at less than this price to get 90% of the benefit" regime--you would no longer de-facto price yourself out of various markets.
(FWIW, my experience on this is that I was a CS person with an algorithms fetish who was recently working with a long-time machine-learning specialist in a weirdly-related-but-not-even-remotely-competing area to your company. The result was lots of conversations both internally and with various of our clients with regards to how to price such a service: the result is that I have seen both a lot of failure modes of advertising and of asking for "percentage of revenue". This is, of course, in addition to running most of my current business on a "percentage of revenue" basis, and seeing the kinds of problems and misunderstandings it can cause.)
I understand where you are going now. To be fair, your point about certain people selling on low margins comes up with us and we usually start those companies at $1/product/month on our pricing and go down from there depending on how many products they want to price with us.
We have been testing this pricing model to see if we will earn more revenue doing this.
I appreciate your continued feedback and your clarification.
We are making all of our vendors more than 1% and for some we've doubled their revenue using a Limited Supply strategy. They weren't a large store but we did make them several thousand more per month.
Pricing is more complex than the classic Econ 101 diagram for maximising revenues under the demand curve on a particular price point. Much, much more complex.
There's also considerations of customer value, segmentation, price fencing, bundling, unbundling, price strategy and so on and so forth.
Of course you're right that pricing is not easy. Your list of other considerations is also very good (sounds like you've taken a course in industrial organization). The only other main idea I want to add is that price is just one of the ways that firms compete with one and other (Bertrand competition). Firms also compete with each other over quantity (Cournot competition), and among other general dimensions like quality (i.e. product differentiation).
In order to have good estimates of price elasticity via a derived demand function, one would also need to consider the price and features of competing products in that demand function (which would enable the use of other economic measures of marginal effects like cross-price elasticity of demand). Your competitor's price is just as important as your own price because of substitution effects.
That book actually has a good discussion on advanced pricing techniques, including taking competitor prices into account. They even outlined techniques for selecting the most efficient bundle of features for a given price -- right up to labs built to perfectly resemble a supermarket.
While pricing is in fact a very complex topic and we actually incorporate some of the aspects of grouping that you are mentioning, when it comes to communicating what we do to small to medium sized ecommerce shops we found that it was best to stick to a more simple approach in our marketing video.
Ha! I loved your technical comment though. It was hard to respond to because I couldn't disagree about how difficult pricing can actually be.
We don't have a lot of marketing people on our team, so our whole idea of marketing is: If we had to explain it to a child how would we do it? Sorry if the approach appears elementary, between the three cofounders we are Math @ Columbia (Me), MS in Software Engineer from USC (CTO) and Stanford Econ/Harvard Law (COO), we don't have a lot of marketing people around so I think we may over compensate and make it too simple sometimes.
I think that in actual fact, your video is good. Lots of small business folk don't even know about the 101 concept of a revenue-maximising point on the demand curve.
I've been staring at these a bit lately because my startup pitch has some demand curves to illustrate what I'm doing.
Technically, the goal of this service (whether the owners describe it this way or not) is to find the unit elastic point on a product's demand curve. In theory, assuming a linear demand function, profits are maximized at the price and quantity associated with the point where price elasticity of demand = 1 (or equally, where marginal revenue = 0). If your current price is higher than the unit elastic price, your total revenue increases by lowering your price. If your current price is lower than the unit elastic price, your total revenue increases by raising your price. These topics are covered in any intro to microeconomics course (as well as any MBA economics course).
That being said, I like the idea, and I wish you guys luck. I'm actually building something similar (YC application in the pipeline) but with a more specific focus on a particular set of markets.
Hi, I'm one of the cofounders. Your description is a great start to what we are doing. There are more considerations that just that but to be fair we aren't publishing papers on economics. We really do want to combine well understood economic principles with machine learning to make a very affordable product for ecommerce stores and event providers.
Hey, I'm Luke. I'm one of the founders of Ventata. This is quite unexpected but I'm trying to bring more people online to help with questions and such.
For now, I'll try to answer as many questions as possible.
19 comments
[ 2.8 ms ] story [ 61.9 ms ] threadMy problem with this is that there is no way that this algorithm is really that impressive to warrant what might be a substantial amount of someone's margin: if you are finding yourself spending a ton of money on this algorithm, you can easily re-implement it... based on both the video and the FAQ it sounds like they are doing simple hill-climbing, which should take a few hours to implement (and which, I will however point out, isn't even anywhere near optimal for this kind of problem).
Also, I have mentioned it before but if you are charged 1% and someone makes you 2%, I don't understand why you wouldn't buy from them. (On a side note: we earn our sellers more than 2% on average)
If I ended up in the situation where I was paying you hundreds, certainly if I was paying you thousands, of dollars a month, I find it impossibly difficult to believe that I couldn't get more "bang for the buck" by spending the time to reimplement the features I am getting from you.
(As an entirely unrelated argument, the optimal margin might not even be 1%: to charge a percentage of revenue for something that doesn't have any scaling terms in proportion to the dollars being moved is, to put it lightly, "bold"; my business, for example, probably couldn't afford 1%, yet I routinely move tens of millions of dollars... and I certainly, at the price of hundreds of thousands of dollars a year, would have no trouble reimplementing your setup, and in fact have done similar things as side projects on various occasions.)
On the other hand, I get it - You would never use us, you would try to build it yourself. I can respect that. I get like that for a lot of stuff I see.
The founding team's past experience is in algorithmic asset valuation, commodity trading and machine learning. We are life long learners and are dedicating our time to find better ways of helping ecommerce stores pick better prices. We have several years of experience in workflow automation and we apply that to our current business to make our product as cheap as humanly possible and pass those prices along. Our closest neighbor in the pricing game charges hundreds of thousands of dollars in fees because they use so much human labor to do things.
Saurik, if you ever want to do one less thing and want us to do your price testing. Please give us call, I would love to find a price that makes it worth your time.
After all, I am fond of dynamic pricing. :)
It isn't because I would be an "enterprise company" that I would need special pricing, it is because your pricing model makes a fundamental assumption about the kind of business model that the other person might be engaged in: if you had a "percentage of profit" model--even though you would still be in the "this is too expensive: I could hire a full time developer and a CS grad student at less than this price to get 90% of the benefit" regime--you would no longer de-facto price yourself out of various markets.
(FWIW, my experience on this is that I was a CS person with an algorithms fetish who was recently working with a long-time machine-learning specialist in a weirdly-related-but-not-even-remotely-competing area to your company. The result was lots of conversations both internally and with various of our clients with regards to how to price such a service: the result is that I have seen both a lot of failure modes of advertising and of asking for "percentage of revenue". This is, of course, in addition to running most of my current business on a "percentage of revenue" basis, and seeing the kinds of problems and misunderstandings it can cause.)
We have been testing this pricing model to see if we will earn more revenue doing this.
I appreciate your continued feedback and your clarification.
There's also considerations of customer value, segmentation, price fencing, bundling, unbundling, price strategy and so on and so forth.
In fact ... whole books are written about it. I read one recently: http://chester.id.au/2012/09/12/review-the-strategy-and-tact...
It's probably the most important business I book I can remember reading. Pricing is the low-hanging fruit, the big juicy lever of profitability.
In order to have good estimates of price elasticity via a derived demand function, one would also need to consider the price and features of competing products in that demand function (which would enable the use of other economic measures of marginal effects like cross-price elasticity of demand). Your competitor's price is just as important as your own price because of substitution effects.
Thanks for the feedback
But I have to show off mah fancy book learning somehow, dammit! :D
We don't have a lot of marketing people on our team, so our whole idea of marketing is: If we had to explain it to a child how would we do it? Sorry if the approach appears elementary, between the three cofounders we are Math @ Columbia (Me), MS in Software Engineer from USC (CTO) and Stanford Econ/Harvard Law (COO), we don't have a lot of marketing people around so I think we may over compensate and make it too simple sometimes.
I've been staring at these a bit lately because my startup pitch has some demand curves to illustrate what I'm doing.
That being said, I like the idea, and I wish you guys luck. I'm actually building something similar (YC application in the pipeline) but with a more specific focus on a particular set of markets.
Thanks for the good wishes!
For now, I'll try to answer as many questions as possible.