What benefit does a Negative Operating Cycle get you?
February 17, 2012 12:19 PM   Subscribe

What benefit does a Negative Operating Cycle get you? Other than just cash on hand. Amazon has a lot of cash because they practice Just In Time inventory management & sell out of their products before the bill for that product comes due. Other companies pay the bill first & then wait for the stock to sell out to recoup their money. Aside from the benefit of having cash on hand, what benefit do you get from this?

I recently stumbled on this slide from this excellent presentation, which led me to this NY Times article about Amazon's Negative Operating Cycle.

This has me wondering - what are the specific benefits? Why were stock investors so excited about this concept?

I have my own conjecture about how this benefits them and I think I'm fairly on the mark- but I want to ask the hive mind to see if there's a better, more accurate answer out there. I'll share my thoughts on this after a bit (I don't want to influence the response too much one way or the other).

Also how realistic is it that Amazon sees money from goods sold in 2 days? I'm not really aware of many payment processors that do that, but if they're their own payment processor then it would make more sense that they could get money in 2 days from Visa etc. directly.
posted by MesoFilter to Work & Money (19 answers total) 4 users marked this as a favorite
Best answer: Well, for one thing it means that Amazon earns interest income on that money. Amazon can literally sell something at zero margin and still make money on the transaction, that is very hard for other companies to compete with.
posted by atrazine at 12:54 PM on February 17, 2012 [3 favorites]

Response by poster: Right, that's what the presentation said. Amazon can make zero margin, or even lose money and still earn money.

a) The interest on e.g. $100 over 25 days is less than e.g. 10% margin they could get selling at at normal prices in a CD or bond.

b) A hidden benefit is that Best Buy has to borrow money to pay for their inventory, so Amazon is earning interest while Best Buy is paying it.

So what else? What do they do with the cash besides put it in a bank?
posted by MesoFilter at 1:33 PM on February 17, 2012

Response by poster: Here are the notes I took on the slide:
amazon can take a loss on every product they sell and still make money, sell everything at cost and still make money

amazon's negative operating cycle

amazon sells every product they have in inventory every 20 days, so they reorder from their suppliers every 20 days.

in retail the average is much higher than that, for example best buy reorders every 74 days - it takes them 74 days to sell out their inventory.

but the standard payment terms are 45 days.

best buy buys their product on day 0 and they pay for it on day 44, and the customers have bought the whole thing by day 74. day 76 is when they actually get the money.

the intervening time between when they've paid and they sold it is debt.

amazon, like best buy orders the product on day 0 and pays for it on day 44 like best buy, but they sold out on day 20 and collect the money on day 22.

that time is called "float" - they're sitting on your cash, so they can give you every product at cost and make money on it.

this is why they give you free shipping, are willing to discount
So the presentation doesn't actually say they earn interest on the money, that's sort of assumed - I suspect the source for the presentation was that NY Times article, which was equally vague about what they do with the float. So what I'm looking for is someone who actually knows what the benefit of a positive cash float is other than "they earn interest on it" and "they have cash on hand instead of in debt."
posted by MesoFilter at 1:51 PM on February 17, 2012

Best answer: There isn't much more to it than what atrazine said. If you make revenue and then pay expenses out of the revenue, you don't need to finance your sales. You hold more cash on average that earns interest, and your backup reserves don't need to be as large. That frees up money for R&D, capital purchases, longer term investments etc.
posted by michaelh at 3:08 PM on February 17, 2012

Best answer: I don't know what you're looking for here other than a) interest and b) cash position. Those pretty much ARE the benefits of collecting faster than you're paying.

Steering somewhat from the strictly financial side of the transaction (it still ultimately goes to the money, but by way of the cost accounting aspect) would be these advantages of ordering as retail orders are received:

- reduced RISK. Obsolete product, product which must be sold at a deep discount because the later and greater thing is announced while you're holding the older and lousier thing (iPad 1's, anyone?).

- a better "futures position" of sorts. Almost all technology goes down in price, sometimes drastically. If you buy TVs that retail for $700, let's say you pay $350 each for them at wholesale. If you're holding them when they drift down to $600, guess what, you still paid $350 each for them. OTOH, if you order them as you go, the sale you make at $600 will be filled with one you paid $300 for...

- reduced warehousing - space, insurance against loss, theft, etc.
posted by randomkeystrike at 3:12 PM on February 17, 2012

Response by poster: Hmmm. Thanks. That's pretty much what I thought but the NY Times article hyped it so much that it seemed that there just HAD to be something I was missing - interest on the cash they have seems to be not enough to actually sell at a loss.

As for my own promised analysis it goes along the lines of this:

Amazon has a large cash reserve which means they have more cash on hand to cover expenses such as invoices from vendors, paychecks, server farms, etc. so it doesn't have to borrow anything (even at a low interest rate) to cover those expenses. During the whole financial meltdown there was much talk of the decreased lending possibly impacting businesses because they couldn't borrow money to cover paychecks - I realized that this is probably what companies like Paychex do - charge a small fee (interest) for doing payroll for you.

So there's a whole host of small fees that simply go away if you're Amazon and not Best Buy.

Amazon doesn't regularly price below their wholesale cost. What they typically do is price just below their wholesale cost + the markup they charge marketplace sellers (listed here). You can assume their profit margins are roughly in line with that chart. If they do price below their wholesale cost it's for predatory pricing reasons, which I have thoughts on but won't speculate on here.

Amazon also has incredible deals with UPS - my rough estimates put it at something like 66% or more depending on the size of the item (smaller items get more of discount) which is part of how they can offer free shipping - because shipping is practically free for them. You can estimate this for yourself by finding an item on Amazon with shipping dimensions listed (not product dimensions), and plugging it into the UPS website to find the Ground Shipping fees and then plugging it into the Amazon FBA calculator.

Users also do buy more when they get free shipping. I read an article years ago (too lazy to dig up the reference) that said that one of the Amazon companies (in France or Germany, I forget) refused to offer free shipping so they offered $1.00 shipping instead, and when they finally offered free shipping sales jumped. Additionally when someone joins Amazon Prime, their annual spend jumps from $400 to $900.

Since Amazon is in a position to lend, rather than borrow money, they can simply lend it internally to fund things like R&D and advertising. That R&D money has paid off in a big way. Since everything Amazon does is a platform, they were able to put much of their technology in the hands of the public and B2B they also make bank selling technology like web hosting, fulfillment, cloud computing, etc.

They also buy up companies - if you read the Acquisitions, Investments & Subsidiaries sectin of their Wikipedia page you'll see how many businesses they're actually in - I was surprised to see DPReview, EngineYard and LivingSocial in that list.

Thanks everyone for your analysis, sorry if I seemed to be asking for the impossible - I really thought there must be more to it than that.
posted by MesoFilter at 4:23 PM on February 17, 2012 [1 favorite]

Best answer: Just in Time ordering also reduces the greatest money pit of all - warehousing excess inventory. You pay to protect it from the elements, you pay to have it schlepped, you pay to have it managed. Buying in quantities you can move quickly eliminates all of this waste.
posted by halfbuckaroo at 3:46 AM on February 18, 2012

Response by poster: I think warehousing is a wash. I doubt operating a 10,000 square foot warehouse is 10x more expensive than operating a 1,000 square foot warehouse and a 100,000 square foot warehouse is certainly not 100x more expensive to operate. You also only have to pay to have it shlepped twice - when you bring it in & when someone buys it.

By buying 3 to 4x as often, they have to pay a larger staff and greater logistics to handle the 3 to 4x more deliveries than Best Buy would have to receive.
posted by MesoFilter at 9:40 AM on February 18, 2012

Best answer: In our case, we went from receiving, delivering and warehousing for fourteen retail locations to zero warehouse and no warehouse employees. That's two supervisors, eight warehouse staff and nine receivers.
That's nineteen salaries gone, and the cost of the warehouse and receiving facilities.

Deliveries now go straight to the retail locations, and they could get 10x the deliveries without adding a single employee.

The savings are significant.
posted by halfbuckaroo at 1:17 PM on February 18, 2012

Response by poster: I was actually thinking about that - if retail locations could get their own deliveries. I was wondering whether or vendors would continue to offer the same kinds of discounts (e.g. "order $10,000 worth of product and get 5% off") if they had to treat them like 15 deliveries rather than 1 delivery.

Also that retail stores needed to keep as many locations as they had stocked where Amazon was centralized into one location - so if you have 15 locations you could do the Just In Time game that Amazon does, but timing the deliveries makes logistics a bit more difficult, etc.

Like say you're Radio Shack and a significant portion of your inventory is random diodes and stuff that people rarely buy, but you need to keep a complete selection in stock for the people that do buy them. If they sell 1 per 6 months then you must keep 6 months worth in stock in each store at all times.

But if you're amazon, you don't sell 1 per 6 months, you sell a few per month so you carry only a few (not that amazon would waste their time with a diode that sold 3 per month but you get the idea).

Thanks for the insight - I agree that going from any warehouse to no warehouse is a significant savings, but in the case of Amazon they can't go to no warehouse since they have no bricks & mortar outlets, unless all of their vendors started drop shipping, so in the case of Amazon - I don't know that per-item, going from 2-3 months of stock to 2-3 weeks of stock is as significant a savings.
posted by MesoFilter at 3:21 PM on February 18, 2012

Best answer: I was actually thinking about that - if retail locations could get their own deliveries. I was wondering whether or vendors would continue to offer the same kinds of discounts (e.g. "order $10,000 worth of product and get 5% off") if they had to treat them like 15 deliveries rather than 1 delivery.

We forecast needs and issue blanket purchase orders for "standard" items for the year, with staggered delivery dates. As long as the vendors know they're going to get the orders, we get the discount for volume regardless how many shipments, even weekly if need be.
posted by halfbuckaroo at 7:13 PM on February 18, 2012

Response by poster: Really? That's interesting. I assume then that the invoices come in staggered as well and you sort of pay as you go?
posted by MesoFilter at 11:23 PM on February 18, 2012

Best answer: Payments are scheduled on the blanket P.O. when it is issued, usually six per year. However, it is a seasonal business, so four payments may come J/F/M/A, when shipments are hot and heavy, with payment five in August, and payment six in December.

Blanket P.O. is also adaptable; if we start burning through a particular item or group of items, the P.O. can be accelerated and completed in six months, and then another blanket p.o. written. Payoff is also accelerated.

I should note that stocking fourteen retail stores requires somewhere on the order of 3,300 vendors. I'm sure that Amazon has several magnitudes more vendors than that, all racing to the bottom on pricing.

Customers the size of Radio Shack and Walgreen's pay all their invoices once a year, or twice at the most. Whatever has been shipped up to forty-five days before pay date gets paid. The rest goes on next pay date.

Any returns/damages/out-of-date products get pulled out of the invoices before payment. That's old school, so you can just imagine what Amazon can demand and get from manufacturers/suppliers. And Wal-Mart? Oy.
posted by halfbuckaroo at 3:27 AM on February 19, 2012

Best answer: ...but in the case of Amazon they can't go to no warehouse since they have no bricks & mortar outlets

So they've chosen to eliminate retail stores. One or the other, but not both.
posted by halfbuckaroo at 3:33 AM on February 19, 2012

Response by poster: Thanks for the information. That's very interesting.

Amazon is very efficient at rejecting non-sellable items - they get returned very quickly. I'm sure they get replaced or refunded efficiently - perhaps tacked on to the next order, or just deducted from the invoice.

I have no idea how many vendors Amazon deals with. I'm sure they also have very sophisticated tools to deal with inventory, demand, placing orders etc.

Do you mind if I ask what sorts of tools you use to forecast demands & deal with so many vendors?
posted by MesoFilter at 6:57 PM on February 19, 2012

Best answer: Not to be evasive, but the software for vendor tracking/invoice/cost analysis/forecast/purchase orders/mail order was a proprietary system written for an IBM AS400 system by an IBM partner, with custom parameters. In 1993. It's still plugging away.

Retail stores used RetailPRO for receiving and POS, with links to the AS400 for real time reporting.
posted by halfbuckaroo at 2:45 AM on February 20, 2012

Response by poster: AH, no that's fine, what we're using is also some sort of proprietary software, though we're moving off of it in the near future. One of the reasons I'm trying to reverse engineer what Amazon does is so I can bake their best practices into what we do, and I think we're at a point where I can (even though I'm new) shape the tools in a positive way - quite a few of the ideas in this thread are things that I'd never heard discussed around the office in any of my training or post training activities.
posted by MesoFilter at 4:45 AM on February 20, 2012

Best answer: Bear in mind that these techniques are most effective for what we call "staples," items that we stock day-in, day-out for years, if not decades. They form about 70% of the stock.

The other 30% is "trending," which comes from trade show orders, vendor visits, etc. These are purchased more traditionally, as in we'll try 30 units and see how it moves. Once we have some sales data on it, we either discount it and move it out, or add it to the staples.

Most importantly, everything, (even staples) gets an annual review for performance/price changes/quality, and lives or dies by that review.

There's a whole nother category for items that are purchased by container load overseas, but they're usually one-offs and tend to become "museum pieces." This is usually less than 2% of inventory, thankfully.
posted by halfbuckaroo at 6:08 AM on February 20, 2012

Response by poster: Thanks for the insights. I agree that a blanket PO couldn't really cover items you had no ability to predict the demand for, though I could imagine a scenario where you have some staple items from a company and some "trending" items and could create a blanket PO that would cover all of your staple items, plus a few different trending items each delivery - changing those items on the next segment of the blanket PO based on how well they work (adding more if they do, removing them if they don't). If that one vendor had staples and items you were interested in testing out.

Sounds like you have a good methodical process in place. I don't think we're that organized, but at the same time, we also have to consistently re-evaluate based on market conditions so every month or two we're re-evaluating each vendor and the product lineup (which also changes continually), so I don't know that as much process can be written around what we do - but as I said, I'm looking for any best practices I can apply to myself & perhaps in some operational way around the other buyers who are amenable to adopting these methods.
posted by MesoFilter at 2:55 PM on February 20, 2012

« Older Suit and tie   |   Automagic help to organising a filesystem? Newer »
This thread is closed to new comments.