How “refreshers” work at Amazon

Last updated: 7/21/2019 TL;DR: If you sign with Amazon, expect only paltry base raises and no new stock grants until the 5th year even if you get promoted. If you don’t, your TC will actually drop to the bottom of band so that’s a cliff in its own way. What you see in the offer letter is all you’re going to get. I constantly see many people weighing offers in here from Amazon and what they see in front of them, they assume it’s just like other companies in terms of annual raises and refreshers. It’s not. Many in Amazon gripe about the total compensation target (TCT) model where it’s possible for you to be rated the equivalent of exceeds and see no difference in pay compared to meets. And that’s because the TCT model takes into account your current grants when determining how many grants you get for the future. When an exceed’s TCT is higher than a meets performer, they both can end up getting 0 stocks for a year because for that year, both their existing grants exceed the TCTs. In order to better understand the TCT model, let’s take the SDE2 role as an example for the purpose of this post. The TCT band ranges around 182-242. Amazon’s fiscal year goes from April to March so when it looks at when stocks vest, stocks vesting in Jan 2019 is part of fiscal year 2018, stocks vesting in Mar 2019 is part of fiscal year 2018, stocks vesting in Apr 2019 is part of fiscal year 2019, Dec 2019 is part of 2019, Jan 2020 is part of 2019, Mar 2020 is part of 2019, Apr 2020 is part of 2020. Let’s say you have accepted an external offer 2 years ago with a TC of $242k with a base of 150k and you have that last stock vest currently worth $100k for fiscal years 2019 and 2020 and 0 for 2021. When you meet for annual review with your manager, Amazon does not give extra grants for 2019 the current year, but you may or may not get some for 2020 or 2021. You performed excellently this year so you’re going to the top of band with a TCT of $242k. Your base is currently $150k, so the extra grants you get is calculated like this for 2020: max(0, (242 - 100 * 1.15 - 150)/1.15) = 0. For 2021, it’s calculated like this: max(0, (242 - 0 * 1.15^2 - 150)/1.15^2) = 69k. Yes, Amazon implies a 15% annual growth rate of the stock when calculating the value of grants. Let’s see what happens when your buddy joined at the same time as you with the same offer, also starting 3rd year about to evaluate grants for his 4th and 5th years except he’s been rated meets so he’s at the bottom of the TCT band. His base is also $150k, so the extra grants he would get is calculated like this for 2020: max(0, (182 - 100 * 1.15 - 150)/1.15) = 0. For 2021, it’s calculated like this: max(0, (182 - 0 * 1.15^2 - 150)/1.15^2) = 24k. Over these 3 years, the comp you will actually get will look like this when calculated in the conventional way assuming zero stock growth: 250k, 250k, 230k. Wait what, wasn’t the last one supposed to be 219k? Well not really because you will have another annual review next year and if the stock fails to grow 15% or higher, you will get extra grants to ‘make you whole’ when calculating as we did above again next year. But let’s see what it looks like for your buddy who just got a meets. 250k, 250k, 178k. Huge cliff when being rated meets. At this point, you and your buddy are only fighting for what you’ll be paid starting 2 years down the line. “Wow this sounds harsh looking at how different the two outcomes are. Are these really the only two outcomes that are possible? Is TCT set either only at the top and bottom of the band?” Yes for the most part, but I did say the most part because there’s a slight nuance. The top 20% receive exceeds, middle 70% receive meets, and bottom 10% is needs improvement. If you’re at the bottom, you will get zero stock grants and Amazon is looking for you to improve quickly or get pipped out. There is a nuance within the meets however. With the meets, there’s a suffix at the end that is enumerated as: with no additional investment, with additional investment. If you get no additional investment which is the vast majority of meets (since the manager has a budget to go by), then your TCT will be at the bottom of the band. If you do get some additional investment, it might go up to 30% range penetration according to the SDM named RBMY63 here, which means target in this example would be (242-182)*0.3 + 182 = 200. He estimates around the top 10% of the middle 70% receive this additional investment. The middle of the band to the top is meant for external hires, while being rated exceeds put you close to the top of the band. People often say they did get a raise in TCT even when being rated meets but that’s most likely because the TCT bands have moved up every year so the bottom moves up, not because they have moved up within the band. Additionally there’s actually a catch that made this TCT scheme even worse this year (2019). It looks like HR changed the formula to screw us even more for stock grants in t+2 (2021 in this case) for meets expectation ratings, internally known has HV (including with additional investment, known as HV+) by suppressing the TCT for that year by ~5%, in addition to the 15% discount every year with the growth rate assumption. This is my speculation based on the numbers I’ve seen from TCT sharing threads, but I’ll update this as soon as someone can provide the actual new formula. Due to this, in the example your stock grant will remain the same but your buddy’s TC with the stock grant will then look like 150 + (182 * 0.95 - 150)/1.15^2 = $167k for t+2 this year with $17k in stocks, and then they may provide additional grant next year at t+1 to make it $178k with an extra $11k in stocks assuming the stock doesn’t grow, but the real amount of course depends on how well the stock does and what the TCT is for that year. My speculation is that HR heard the complaints about receiving no new stocks for a particular year loud and clear so they’ve decided to screw us even more by giving less for t+2 increasing the chances we’ll get some stocks next year for t+1 to give the perception that Amazon now gives refreshers every year. Of course it works nicely for them because the TCT is actually suppressed by 5% in case the stock grows more to the point that it exceeds TCT. So we’re screwed even more if the stock does well since they won’t be giving new grants for t+1 whether they give us 95% TCT for year t+2 or 100%. For exceeds expectation ratings, it seems like the formula for t+2 remains the same. “Okay, this sounds definitely like they’re out to screw us, but I still want to stay at Amazon because I’m too lazy to leetcode. What if I get promoted? Won’t I see a good raise there?” You’re probably not seeing much shit either. The difference in TCT is minuscule because you’re likely going to be rated exceeds for the SDE2 role and be put to bottom of the band for SDE3 afterwards with the meets rating as you have a new bar to be evaluated on. The SDE3 band would be something like 250-370 or so. Your TCT goes up from 242 to 250, and all the stock grants you may get is offset by your base increase from 150 to 160. That’s your raise for being promoted, that’s it but at least you’re staying there instead of taking the risk of snapping to the bottom of the SDE2 band. “Wow they’re really screwing us, but it can’t be that much better at other companies right? Don’t they also have a cliff?” Yes they do, but that’s only because they actually give refreshers on top of the initial grant. Let’s take Google L4 for example. You’re offered $240k TC with a $150k base, 15% bonus, and $270k in stocks. Let’s say refreshers are $80k a year, these are all reasonable numbers for meets. Your first year TC is $240k, but due to annual refreshers it’s $260k the second year, $280k the third year, and $300k the fourth assuming no promos/raises or stock growth. Then the cliff comes and you’re left with a TC of $230k. Compare that with the previous Amazon scenario and you’ll clearly see the numbers being higher here. “But Jeff Bezos told me Amazon is a long-term company to work for! My manager also told me the great thing about Amazon is that there is no cliff once I get promoted.” It’s very funny based on how Amazon still manages to call itself a long term company, at least when it comes to compensation. Over the long term, we’re getting screwed because while external offers can possibly be slightly lower due to our stock going up faster than other firms in the recent past, the fact of the matter is that over the longer run growth rates have been similar. Then instead of having our pay indirectly suppressed over the long run with the TCT model, having additional refreshers that stack will in fact be better for us in the long run. Your manager is also correct, you won’t have a cliff if you get promoted. There is no cliff because there were no refreshers to build the cliff in the first place. *points at head* “What about when the stock goes down? Because the TCT means the gap is always filled, we get made whole right?” Have you seen how Amazon treats its fulfillment workers? Not from their propaganda videos but from actual accounts of working there? If/When the stock goes down, it’s very likely we’re in a recession. Amazon’s the type of company to tell us we’re lucky to have a job when that happens so they’ll probably end up lowering the TCT bands for that year for ‘tough market conditions’ despite the fact that RSUs don’t impact cash flow. And yeah, that might not even matter if you get laid off. The other flaw is that bear markets are simply much shorter than bull markets. Amazon may win for 1 or if lucky 2 years, but there’s the other 6-10 years of bull market where the traditional model is favored. “Okay, I see Amazon’s playing a game of TC trickery by luring us with offers that look competitive on the outside but screw us once we’re in. But I don’t have offers from other FAANGs, so should I join?” When you evaluate an offer, look at what your TC will likely be over 4 years compared to getting one from another company, and the worst case beyond that and then make your decision. I also recommend current Amazonians looking for external roles to do a 4-year evaluation of TC when comparing offers of staying in Amazon vs jumping by adding refreshers from that other company by using information here in Blind as well as accounting for the 15% premium for each year forward in stock valuation for Amazon. As we saw above, Amazon actually gives you less than your TCT in grants because it assumes a stock growth rate of 15% every year. It also helps to see what the pay looks like beyond 4 years to get a worst case scenario outlook as well if you don’t get promoted or unable to find another job by then. Additionally, you can look at the “signing bonus” as part of the stock grant converted into cash instead. So instead of a 25/25/25/25 vesting schedule with the stock, it becomes 3.125/9.375/25/25 and the signing bonus is supposed to be 21.875/15.625/0/0. I made it say 25 at the end and not 40 because you’ll notice that compared to similar offers at other companies, the overall stock grant is actually lower. Then there is really no signing bonus or annual bonuses to top that unlike other companies. Another thing I mentioned above: If your offer was at top of band, which it likely is for external hires and you have been rated meets for over 4 years and not gotten promoted, your TC will actually DROP to the bottom of the band. Then your TC will be a lot lower than what you have first received. That’s a lot of text but I’ll be keeping this in my post history as long as I can so I can link to this (and suggest others to do the same) whenever someone received an Amazon offer and is weighing the options. Also I made this post to raise awareness among the tech community of Amazon’s cunning compensation practices once they are inside so that engineers can make more informed decisions on whether or not to join Amazon or leave if they’re inside. Please feel free to copy and paste some of the things I wrote on Reddit, Hacker News, etc.