Why does Price (not) matter in data? | Hacker News Reader

Why does Price (not) matter in data?

medium.com

2 pointsdaehee11 comments

These kinds of articles always make me a little sad. They trick me. I start to believe “rticle” follows the A, like a good language model, and in my horror I find it should have been “dvertisement.”

> Spending too much on snowflake? This one tool (of ours) can help (for some cash) reduce how much cash you are spending!

Spend to spend less. For a company struggling with spend, perhaps that’s the argument of a fixed price. Think less.

Fixed vs usage based isn’t really important beyond the bottom line. Data, Serverless, doesn’t matter. It’s just a means of charging dollars, and as long as the right amount of dollars are spent, all is well.

Shouldn’t be important at all without sufficient accompanying headcount. What good is a cost-effective tool without a hand to hold it? Funny thing, that.

I think we should focus on the key topic: tackling high costs. While the backdrop for the article is content marketing for iomete, the point is a higher-level discussion on how companies can optimize their spending on expensive tools that we take for granted, e.g. Snowflake.

Plus, the "price as a weapon" concept is worth exploring further - for aspiring category disruptors in this current economic cycle. It would lead to more competitive pricing and benefits for the customer in the long run.

Just ran your comments - not just this comment - through ChatGPT and am passing along the feedback: As an AI language model, I do not have personal opinions or emotions. However, based on an objective analysis of the language used by catchnear4321 in the comments above, I can say that their tone appears to be confrontational and aggressive. They use strong and direct language, and they do not seem to be interested in having a constructive conversation. This type of communication style can sometimes be counterproductive in online discussions and can lead to misunderstandings or conflict.

you shouldn’t submit the work of others as your own

also ChatGPT said you lied about me.

and that you smell of cheese.

LOL!

If demand for database products is/was inelastic, why are so many SaaS data vendors not profitable after IPO? (Elastic NV [0], Confluent [1], Snowflake [2], ...) They should just be able to increase their prices to become profitable. But they don't, ergo must be competing on price.

[0] https://finance.yahoo.com/quote/ESTC

[1] https://finance.yahoo.com/quote/CFLT

[2] https://finance.yahoo.com/quote/SNOW

The article doesn't claim that demand (or pricing) is inelastic in the data space. It makes the point that VC funding has inflated price levels across the board, similar to how college tuitions rose dramatically after student debt became widely available. There are still differences between college tuitions, but overall tuitions might be multiple(s) higher compared to what the price would have been if there was no (cheap) financing available even.

Now that VC funding is mostly on the sidelines the question is how this will play out further.

Haven't studied Elastic and Confluent, but Snowflake seems now profitable (5%) with 25% free cash flow. See p.28 of their last investor presentation. https://s26.q4cdn.com/463892824/files/doc_financials/2023/q4...

> It makes the point that VC funding has inflated price levels across the board, similar to how college tuitions rose dramatically after student debt became widely available.

What's the evidence for that assertion? For the analogy to work VC funding would need to be a dominant factor in IT spending. Last year it was about $200B vs ca. $2T of overall IT spending. [0, 1] But VC spending is not IT spending--in fact it's probably on the order of 10% because most of the money goes to things like R&D and customer acquisition. So it's fair to argue VC spending has inflated salaries but it does not seem like enough to move the needle on IT spending in a material way.

I'm personally skeptical of the argument from the cited article because I run a company that does pretty much exactly what IOMETE is proposing. In practice, there are many confounding factors.

1. The database market is very competitive. There are very few segments that don't have 2, 3, or more substitutable products for green field applications. For existing application there are high switching costs that mute impact of lower product prices. In plain English: we lose those deals.

2. We've seen impact from startups losing funding, but it does not affect prices so much as revenue when customers simply disappear. The same thing happened at the end of the Internet boom in 2001/2.

What will make a difference in this market is offering a completely different model, such as fully open source projects that promote competitive vendor offerings without high lock-in.

p.s., Thanks for the Snowflake numbers. I just went off their last reported results.

[0] https://www.ey.com/en_us/growth/venture-capital/q4-2022-vent...

[1] https://www.statista.com/statistics/821769/us-spending-it-pr...

It's not clear what the value add of this is vs OSS DIY Spark + Iceberg?

Free option is already available through OSS Spark/super low cost EMR. What's your value add over those?

Aren't the switching costs the issue with large customers.

Not necessarily. Switching cost is an issue when the effort to switch is bigger than the potential benefits of switching (where benefit can be added functionality or lower costs).

For instance, average Snowflake bill is $300k and many Snowflake customers (almost 400) pay more than a million per year. If you can cut that in half by augmenting Snowflake with IOMETE that's a pretty sweet deal.

No need to switch (that's a hard sell for an early stage startup to enterprise customers), but just augment...

How? By transitioning some of the compute load to IOMETE data lake. IOMETE charges a flat fee compared to Snowflake's consumption-based model. By cutting Snowflake compute consumption (often 5x or more the price of an AWS instance) organizations can save a lot.

Reply on news.ycombinator.com