It's nice to have the right estimate. It's nicer to have it for the right reasons. Most military campaigns these days undergo an "after action analysis." On Thursday March 2, Micron Technology's (NASDAQ:MU) CFO went into battle at a Morgan Stanley conference. We need an after action analysis after yesterday's pre-release, as this article's title implies. What are the lessons learned? Where were we right? Where were we wrong? What new information can we employ next time around? Not only did the sell side analysts do an abysmally poor job with their consensus estimates of 66 cents, vs. the 86 cents now guided. But they haven't told us where their analysis went off the rails or given us confidence they're going to do a better job next time around.
For my part, I'm delighted to average my 77-cent base case estimate and my 95-cent high side estimate and announce that I was spot on. But it would be better to have the right estimate for the right reasons. Here's a little synopsis:
|EP Base Case||$4.746 Billion||35.8%||77 cents|
|EP High Case||$4.93 Billion||38.3%||95 cents|
|3/2 re-guide||$4.65 Billion||37.5%||86 cents|
As with most of the analysts cited below, my revenue forecast was off significantly. In my case it is almost certainly bits for both NAND and DRAM, but it is also almost certainly price. This will be a key detail to look at after the formal earnings call on March 23. If the ASP increase was significantly below the 15%/1% I used for DRAM/NAND, I will be looking to see if this will then bleed over to the next quarter, with a lagged positive ASP effect from this February quarter. What has the company done to so significantly increase their gross margins? Rather than some of the inane and unanswerable questions analysts ask on the earnings call, this is the sort of thing the analysts should be drilling in on. If Micron has been able to carve out a sustainable margin improvement, god forbid, this might entitle the company to a healthier price earnings ratio than our maximum 10x.
What about inventory? Readers of my estimate articles know that I'm interested in the effect inventory is having on reported EPS. Much of the $2.75 billion in inventory that was on the books as of the end of the first quarter was put there at a much lower price. When it is sold, that difference in price will flow to the earnings per share. Plus, Ernie warned us he is planning to do inventory liquidations in this analyst day slide:
And here's what the analysts that bother to show inventory showed in their post pre-release notes:
Cowen and Credit Suisse: You get an A+ for listening during analyst day. Goldman, Needham and Citi: please write us a report telling us why you don't believe Ernie when he tells us he's going to do inventory liquidations - you may have very good reasons and we'd like to hear them. And Citi, as the high entry you should get prepared for detention. Goldman and Citi: please follow course instructions and deliver your models on a quarterly and not just an annual basis. Thank you.
And what about revised estimates? I grade Investor Relations very hard. If I were Ivan Donaldson, VP IR for Micron, I would be giving myself failing grades looking at the estimate dispersions below, coming on the heels of a pre-release guide and a recent analyst day:
Earnings per share estimates:
Wells Fargo and Citi: Do you really think Ernie is not going to deliver at least the 86 cents from the pre-release guide? Are you aware that the last two quarters, the actual was even better than the pre-release?
And how about revised revenue estimates? Here again, there's a big dispersion:
Charts! What about some charts? Readers know I like charts and even that I try to keep them updated roughly weekly on Instablog. And indeed there are some doozies in this bunch of analyst upgrades.
Imitation is the sincerest form of flattery, so here's one from Credit Suisse I found flattering:
Memo to John Pitzer & Co at Credit Suisse. Try normalizing the data as my two charts below do. I think it helps. Also, as my Instablog update link above says, its no longer just about DDR3 as more and more bits are moving to DDR4. Ernie made this very clear in his remarks during the pre-release. In fact, most of the flat to down move since February 10 has been in DDR3. Here are my two charts attacking basically the same data:
And finally, here are two I found a bit bizarre on the same topic. First from Goldman Sachs:
And second, in the same vein, from Cowen:
I guess conceptually I'm scratching my head. I'm delighted with the implication that with a new-found robust gross margin the stock is about to fly. But it doesn't make much predictive sense. For a company that is just qualifying and rolling out a 1x node in DRAM, a 64 layer NAND, and the brand new 3DXpoint, initial margins on all of those should be expected to be poor. And yet the stock could be flying on expectations. Also in the news, management could elect to do a bonehead deal with Toshiba on a solo basis, instead of joining with Silverlake/Dell/EMC, Apple (NASDAQ:AAPL) and Microsoft (NASDAQ:MSFT) as I would prefer. And in that case, even with soaring gross margins, the stock could plummet. I would submit that gross margin is a flawed predictor of stock price despite the tight tango the two lines on the charts above are dancing.
Conclusions: To the sell side analysts: excuse my snarky tone. I know full well how difficult it is to parse through the Micron information. It would be helpful if you could provide a bit of introspection on where your consensus estimates of 66 cents went afoul.
To the SA readers: Thanks for the many positive comments on my aggressive earnings predictions. I consider myself lucky and foolish for having listed both a number and a date for pre-release. And, as stated upfront, it's nice to be correct on the number, but it would be nicer still to have all the underpinnings correct - they weren't. The process of forecasting the next quarter can't begin until we have the detail customarily unveiled during an earnings call.
To Ernie: Everything is not a secret. You can tell us poor shareholders things your customers, vendors and competitors know within a gnat's eyelash of accuracy. Please tell us on earnings day how long your average contract is, what percentage of your production is presently under contract, and the percentage of production going into the spot market vis a vis a long term average of selling on spot. Please tell us how you are doing on your inventory liquidations you said were being undertaken. And if they haven't begun, why? Oh, and I'm sure you will have some pre-payments for product as in years past. Are these all non interest bearing? Are they non concealable? How many customers have done this? What's the average duration? You said on 3/2/2017 that DDR3 was increasingly unimportant. What's the percentage of DDR3 compared to DDR4? I don't think revealing any of this would put you at a competitive disadvantage. Treat your analysts and shareholders better and they will reward you with a higher peak PE than the measly 10X we've seen recently.
I'm feeling we will publish $1+ in EPS and $5 billion-plus in sales for the May quarter. But we can't begin that star gazing until we have more spot and price action, and the data from the earnings call on March 23. Good luck to all!
Disclosure: I am/we are long MU.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.