Valuing shares of Amazon (AMZN) has always been a difficult task since the company does not at all care about short-term profit margins. Investors are left with trying to estimate, based on the company’s various businesses, how large each will get and what type of margins will likely be achieved once each reaches maturity.

Of course, such an approach becomes nearly impossible when you have no sense of which businesses Amazon will choose to enter over time (or maybe the better question is which they will “not” enter). Traditional retail was one thing, but now with cloud services and advertising revenue, margins are going to be all over the map.

I recently trimmed many of my clients’ positions, as I have done once or twice since I made the investments beginning in 2014. My methodology has been inexact, to account for the aforementioned issues regarding Amazon’s various ventures, but it generally involves looking at AMZN on a price-to-sales basis and then seeing what margin/multiple assumptions are baked into such valuations. For instance, if you think they will ultimately earn a 10% profit margin at maturity, you might be willing to pay 20 or 30 times normalized profits, which would equate to 2-3 times annual revenue.

Given the company’s growth, my personal view is that anything up to 3x revenue is at least somewhat reasonable, as I don’t see margins going above 10% given the company’s desire to remain value-based in the eyes of consumers, and anything over 30x profits for a growth company makes me nervous. And if someone argued that they will never reach 10% margins and a 30x multiple is too high, I can totally understand that view. I just think some valuation flexibility is warranted given that Bezos might actually get as close as anyone in business to total world domination.

So below I have posted updated graphs that show Amazon’s stock price over the last two decades or so (not very helpful when trying to evaluate the valuation), as well as their year-end price to trailing 12-month revenue ratio (far more helpful in doing so). Note: the 2018 data points are based on today’s stock price and consensus 2018 sales estimates.

As you can see, AMZN stock hovered around the 2.0x price-to-sales ratio level between 2004 and 2014, with a range of 1.5x-2.5x or so. In recent years, as momentum stocks have led the market higher, that number has surpassed 3x and currently sits around 3.2x.

Given that position sizing in portfolios is always important to me, this graph tells me that now is not a bad time to trim AMZN. Trading above 3x revenue would seem to indicate that investor sentiment is quite high. There may be good reasons for that, of course, but as the company gets bigger and bigger, its growth rate is sure to slow. In fact, in order to grow by the 29% rate that Wall Street analysts are expecting in 2018, total sales need to rise by a stunning $51.5 billion. That very well might happen (and acquisitions like Whole Foods will only help), but when growth slows to only 10 or 15%, investors might not want to pay 3.2x revenue any longer. In my mind, anything above 3.0x tells me to tread carefully.

What do you all think? What kind of profit margins do you think AMZN will earn at maturity (i.e. when its growth rate is in-line with the average company)? What multiple of revenue seems right to you?