Datadog “optimization”

In the case of Coinbase, building in-house following a $65M bill was a clear no-brainer. They could hire a team of 10 senior and staff-level engineers in the Bay Area, and still have this team cost less than $5M/year. And they then only need to budget for the infra costs, which they can presumably bring down to low double digits per year.

Coinbase planned to move off Datadog, but ended up staying. However, it is not the only larger tech company thinking about bringing observability in house. I have another exclusive which even Datadog might not be aware of, yet. This report is about Shopify and its plan to move off Datadog. But could recent layoffs change things?

I found this interesting in the case of DDOG. Optimizations for big spenders really just sounds like “threaten to leave and churn, but stay after receiving a big discount for future bills”. One has to wonder if this applies to companies like SNOW just as significantly — or, perhaps less so, because using SNOW is more likely to generate revenue for the customer — it’s not all about indirect cost savings like DDOG


This is very interesting…
I would think spend with DDOG also associates with revenue generation for customers… e.g. Coinbase app needs to be monitored for performance… So its not a back office cost… ideally, like MDB, DDOG should figure out how to get their bill out of COGS rather than G&A cost line on their customer…

To your point on SNOW - they already proactively reduce price each year… they generate bigger bill like AWS… and spend with them probably falls into both back office spend and revenue generating activities…

nonetheless, i think both DDOG and SNOW are very savvy and aware of “more than fair share” they pull out of their customers…

now Cloudflare is opposite extreme… with Prince mindset of disrupting from the bottom with lower prices, they seem to have culture of not charging enough… if they can get to half the “value extraction” of DDOG / SNOW, they would exceed rest of these companies both in revenue growth and also in cash generation…


Wow. Thank you for sharing @jonwayne235 This was the most interesting paragraph I found in the blog post:

I asked an engineer at Coinbase who used the in-house stack and Datadog how they felt about the decision to stay on Datadog. They said it was ultimately the right decision, considering the reasonable costs, and the superior Datadog development experience.

Even after “ironing out the issues” and double build with the Grafana stack for months - the engineers still come back to Datadog and say it was “a superior development experience.” Quite eye opening when you see a build vs buy decision playing out like this because developers often strongly prefer build over buy as a matter of pride. Somehow Datadog was still able to win the devs over in this case.

In any case, a 65M bill for observability is insane and means the control inside of the company (Coinbase) was likely very lax at the height of their IPO. Same thing happened at many other companies and I don’t see much of this kind of “low hanging optimization” going forward.


I can confirm Airbnb churned from Datadog in 2020 and moved to in house solution. (This was probably a primary part of the ‘optimization’ Datadog mentioned in 2020. That’s the trend for a lot of silicon valley companies considering the ridiculously high price of Datadog. I’m working for another big tech company and we use in house solution built on top of Grafana. IMO, Grafana is huge threat to Datadog’s growth, especially in tech industry.



Interesting, as Aribnb was a customer adoption highlight for DataDog. Actually still is there on DataDog’s website:


My information source was an engineer who works for Airbnb. Apparently, cutting Datadog and moving to in house solutions was one of the projects they took to survive at the beginning of pandemic. Not sure what is the latest status though.


Just to be clear, @monkeydluffy, I wasn’t doubting you at all, I was pointing out that things have changed in the last 7 or so years, with Airbnb being a long-time and once reference customer, now apparently moving on from DataDog. It’s an interesting development.