From an engineering perspective, this makes perfect sense!
But management reacts negatively, maybe even becoming skeptical of your value to the org. Why?
1. The savings you propose are too small to make a significant difference to profitability. $100k is equivalent to between .3 to 1 FTE engineers depending on geography.
2. Someone has to run this project, they will cost money. Engineers will have to be involved, they will cost money. Every project team is going to have to evaluate their usage and provide justification documents, so now the costs are cascading through the org.
3. Remember that the cost of a decision is the sum of the the cost of the thing done plus the cost of the things forgone. People working on this are not working on things customers will pay for—that is, things that get multiplied by many customers now and into the future and differentiates from the competition. This project has its own costs. The savings must be very, very high to compensate for both.
4. It’s hard to imagine how the savings can be sustained without significant new bureaucracy. Every resource allocation is going to have to go through an architecture, implementation, harmonization, and business necessity justification review. That is a lot of new gate keepers.
5. You indicated that teams already “can’t/won’t/don’t communicate.” This means the teams are going to have significant political battles over who owns/runs which resources. Ownership of a the architecture is a tool to bludgeon other departments into compliance—which will make inter-team conflicts worse because they can’t just agree to disagree. Suddenly, one team winning an argument means winning forever.
6. Solving the previous point with better communication is not viable—it would mean solving fundamental problems in organizational management and psychology.
So, what can you do? The fundamental issue here is that the scale of the problem does not match the scale of the organization. Cost saving projects that succeed—like google early on deciding to develop their own servers and racks instead of buying off the shelf—provide a sustaining competitive advantage. Google saves so much money from its infrastructure investments that it lets them build products that would be wildly expensive for others to replicate.
In business language this is referred to as cost leadership. Ferrari has a much lower need for cost leadership than Toyota. Ferrari still needs to contain costs to be profitable, but it primarily competes on differentiated products in a focused market. They might be interested in saving money on aerodynamic simulations so they could do even more of them. But Toyota would quickly go bankrupt if they didn’t make cost leadership part of everything they do—just a few years a failing to improve efficiency would lead to their cars costing many thousands of dollars more than competitors in a market sensitive to price.
Which kind of company do you work for: one that serves the broad market (AWS) or something more niche (IBM Mainframes)? One that competes on price (commercial airlines) or one that competes on differentiated features (private jets)? Craft your project proposals to the business and they may have a much higher probability to getting heard.
It's the type of reply I want to write every time I see misguided comments from people who don't have the full picture of the business and don't even know that they don't have the full picture.
> Where I work, we're basically donating money to the Bezos charity known as AWS at this point.
And ended with:
> And I know we're not alone and that there are other places (even in my immediate geographical area) that have waste well in excess of what we're pissing away.
That was my point. I perfectly (okay, maybe not perfectly, but well enough to understand that it's not worth my time to pursue) understand why and how these things happen, it isn't my first rodeo. And, honestly, I think I'd probably applaud the guy for figuring out how to get _damn near every one in SV_ to open up their fat fucking VC wallets to the man if it weren't for his business practices, like what we see in TFA.
You do sound resentful, though. Is it because you think an aws customer has a moral obligation to divert money they would save in aws costs into salaries for their own employees as opposed to finding aws employees so that bezos doesn't get his cut?