How do you guys feel about products that competes on price?
Hi Everyone,
Just thinking out loud here and trying to get some feedback. I am a backend developer and was thinking of some products to make. Im thinking of some products that are not unique but can compete on price. Using services like AWS Lambda for certain tasks can be dramatically lower than running a regular instance. Like 1000x cheaper. I know pricing is not everything and wanted to get people thoughts about this.
Give you more context, im thinking of products like Monitoring. A product like Pingdom and a few adjacent verticals to it where I can come into the market with the same product but with a much lower price for the same product.
Any gotchas here? Any great case study for or against competing basically just on price alone for the same service?
"The problem with the race to the bottom is that you might win.
You might make a few more bucks for now, but not for long and not with pride. Someone will always find a way to be cheaper or more brutal than you.
The race to the top makes more sense to me. The race to the top is focused on design and respect and dignity and guts and innovation and sustainability and yes, generosity when it might be easier to be selfish. It's also risky, filled with difficult technical and emotional hurdles, and requires patience and effort and insight. The race to the top is the long-term path with the desirable outcome."
— Seth Godin, The Race to the Bottom
https://seths.blog/2012/08/the-race-to-the-bottom/
Never understood the hype around Seth Godin, all his books could be titled "stating the obvious". And while the race to the top makes more sense to him, his content is far from being top quality.
Building a business around low prices is the oldest strategy in the world and worked pretty well for Walmart, Amazon, Costco, H&M, Zara and many others that are making more than a few bucks and have been around for a long time.
By the way, I am not a fan of low prices because I don't like to deal with price-sensitive customers, bu that's a personal taste that says nothing about the strategy itself.
All of the brands you cite are selling physical goods. It's much harder to establish a defensible position with intangible offerings.
you can think about all the email marketing based on Amazon SES like EmailOctopus, domain and hosting providers like Namesilo and Namecheap, live chat tools like crisp and tawk.to, online design tools like crello and stencil, lead generation tools like snovio. I would argue that form a business side you can add Facebook and Instagram to this list because their ads have the lowest CPC on the market.
I don't think that's harder overall. First of all, if your prices can be significantly cut down you have an inefficient structure of cost- Second, you should be able to work on the quality of your product and get to know customers better than the competition by the time they try to cut down your prices.
these tools are competing on features not price
Amazon SES is largely competing on price. It's also growing like crazy and is an infrastructure layer within the offering of EmailOctopus and others.
You have a good point and thanks for the blog link. I am doing calculations on how many customers I would need and I would need a whole lot more to make enough to justify doing this business.
However, don't you think the cloud providers (AWS, Google Cloud, Azure) are partly racing to the bottom? They have products that are on the high end but the foundational items like compute, storage, and networking, they are all racing to the bottom. Lowering prices one after another to compete.
Im not agree with that but just making a counter point.
I am also seeing this and in some ways, im sacrificing on some parts like design and functionality but as technology moves forwards there are techniques (which i briefly described above) that drive the cost down. Should I not take advantage of this and pass it onto users?
Cloud providers costs are driven by cost of electricity and real estate. They are also layering higher cost service options on top of their basic offerings.
It sounds like you are not convinced competing primarily on price (but not solely as there will be value added features which the competitors does not have. Ive described some in the posts below.) will not work?
I mean im not totally convinced competing on price is the correct way to go about it and that is why I am posing this question to the group.
Specially for a product like monitoring, there are already plenty of Free options out there. One example is uptimerobot.com (a happy user). They don't charge you anything for first 50 monitors. So you have to pay your customers to leave them for example. (obviously not realistic).
Now, to your broader question. I would say that competing JUST on price won't work. You have to solve a problem for the target customer. Perhaps one of those problems could be that a competitor's product costs a whole lot. Ok then you have to ensure that you create a product that still has everything your competitor has but provides more value through lower cost. Remember for software, it is not just about Price. A real business would never just buy something due to lower price. They look for the product and the underlying company's ability to deliver the goods which includes good support, uptime, maintenance, customizations etc.
Another challenge and a big one for me is how to get paying users. I totally agree with you that there has to be some reason for a user to change service especially once they got it working already. Just the development cost and time to change for a service that is free or like $50/month probably wouldnt be worth it.
I didnt fully lay out my idea above. Let me add to it to see what you think of the value and differentiation. Anyone can use this service, it will just be a web GUI like uptimerobot.com and an API. However, I am thinking the first initial target set of users would be Kubernetes users. We will build an operator for Kubernetes that is part open source hooking into Prometheus and an opt in paid for option. If you want to locally monitor endpoints and put the metrics into Prometheus, cool. If you want to monitor your endpoints from around the world, then that is a paid service from us. This "operator" will do most of the work for you. Because Kubernetes has all of the information about your public endpoints the operator can discovery this automatically. So the operator can discover all of the endpoints without any configuration from the end user and monitor them as new one gets added or remove old endpoints as they are removed from the system.
Im thinking Kubernetes users is a good first set of users because Kubernetes is new and people setting up these systems might not have something in place. Although, a lot of people are migrating to Kubernetes so they would probably be using something already.
However, the special bit to this is the "operator" that automatically discovers and configures the endpoint monitoring.
What do you think?
You say 1000x cheaper. How much in $ will that be for your target customers? Only if we're talking thousands of $ cheaper would it be appealing, in which case you stand no chance in the race to the bottom vs bigger players.
In software, I'm convinced it has to be about something other than price. Convenience, targeting, distribution or a very specific feature are the way to go.
It will be like 10x cheaper for the target customers.
If that's true, go for it. I am not a fan of price sensitive customers, but if you are able to do something decent at 1/10 of the price, you can be onto something.
I disagree there. AWS Lambda and "serverless" services tend to cost far more.
See: http://einaregilsson.com/serverless-15-percent-slower-and-eight-times-more-expensive (>1700 upvotes on HN!)
I am a DevOps engineer during my day job and have went through many cost analysis for clients of mine about serverless versus regular instance costs. If you compare a regular server running 24/7 to a serverless setup running at the same rate, then yes. It is more expensive. We did a report a few months ago and serverless cost for CPU and Mem was 5-10x more than the cost of CPU/Mem for regular instances.
Our conclusion was that serverless is not for every type of workload. If you did find the correct type of workload it can be significantly cheaper. For example, the one I pointed out. Monitoring like Pingdom. For one endpoint you make a call 1 time every min for less than 20 seconds. The cost of that is something like $0.0003 cents. That type of workload also requires that you initiate the call from different places around the world. This type of workload is a lot cheaper on a serverless setup than with regular servers.
What im coming down to is that I am seeing a real cost saving with some type of workload on Lambda/serverless type infrastructure. Where this can change the cost structure for a service. Is "monitoring" the correct product to apply it to? Im not sure. I suppose I have a solution looking for a problem here.
Happy to work with people if they think this is interesting.
What makes you so sure? It sounds to me like the issue is just that the task is such a ridiculously small workload that you don't want a whole server. It's so small that it fits in the free tier of some services, though. At the core, "serverless" is all powered by servers, there's an additional layer of software and the provider has to be making a profit at some point.
From my POV, if I can support a million monthly users on a $5 or $10 per month VPS, even if lambda functions could reduce that cost, it's just not worth setting up.
From your POV, if you're selling something for fractions of a cent, how can you even bill for it without payment processors taking more than the entire transaction in fees? If you're going to cut someone's costs, I'd try to look for something where you can save them at least $50, and where you can hopefully capture $10 for your service. Ideally it would be a scenario where both those numbers are much larger!
@ybrickman made a comment that seems partially similar to this comment. If going for the low amount like $1/per-user/month to get $1k MRR I would need like a 1000 customers versus a higher amount say $100/per-user/month I would only need 10 users.
Makes sense to me that it is a different kind of game there. Which one is easier? Not really sure.
You do bring up a good point about the payment processor that will need to take a fixed percentage and the lower it is the higher percentage of the amount it would be.
Depends on what you want to achieve. If you want to build a long-term business then you should concentrate on creating a valuable product/service... and this will be difficult with low prices. Especially in the beginning when you have to invest in R&D.
However, if you want to offer something for free (or cheap) just to build a client base and then sell them something else then it's OK to compete with others on price even if your product is not as good as theirs.
It really depends on your long term goals!
Thanks for the comment. You make a good point about maybe being able to build a customer base here and then sell them something else. I haven't thought of that. I however, thought of this as the first service and then add on to it with additional services. Similar but not put as elegantly as you did.
Ill have to think about this more but thinking about two products down the road when i don't have one feels like a tough job for me.
Almost 15 years ago I started with 2 products each about $20. They helped me start building a client base and step by step my company now has over 70 products (average price $100) and a few SaaS. However, there are many people who just enjoy building products and giving them away for free under the Open Source licence - they do not want to make any profit but just to satisfy others and themself :)
So like I said it really depends where you want to be 5 years from now!
I think im in the camp of trying to monetize on the products.
It sounds like I want to be where you are 5 years from now =)
Wow, I have so many questions for you!
How do you juggle 70 products?
What was your first product?
How do you select which product to build and how do you validate these products?
:)
You need a great team to manage the number of products we currently have. After product 5, I started hiring people and slowly built an organization which now works almost by itself.
I would say that I picked the products by intuition, but the truth is that 15 years ago the competition was a lot less and no matter what product you create, it will sell. I've never waited too much before releasing a new product - as soon as we have something working, we launch it. And if clients like it we start to improve it, if they don't, we drop it. Even now we have products which need improvement, so if you wait for the final version before launch, it will never happen :)
@VEVS How much time would you spend on a product before launching it? Is it 1 week or like 1 month type of thing?
On average it's about 45 days from the idea to version 1.
This comment was deleted 7 years ago
Getting 10 customers versus getting 1000 customers to get to $1k MRR makes sense to me. Putting it like that, I agree that getting 10 customers seems easier.
This gives me a lot to think about. Im not totally convinced that lowering the price is the way to go.
Again, if the pricing is 1/10 and the quality decent, you won't have problem finding 1000 customers. There lots of email marketing services that did the same with Amazon Ses. And they didn't lower the pricing that much.
Interesting point about AWS SES.
Im don't intend to create a subpar product. It might lack some features that the higher price ones have (initially) but this would be a quality type of product.
Encouraging and good feedback.
However, I am not thinking how the heck would I get 1000 customers. Getting customers is probably one of the areas where I am the weakest.