Microsoft Yahoo Merger

The Chickens Come Home to Roost

Here is my theory, just based on personal experience, with no hard facts (but some rationales) to back it up:

Companies know more about their revenue sources than they are required to report. FASB rules require them to be consistent about how they report, but they are also allowed to change how they report from time to time. You can "book" revenue from a sale immediately, or spread it out over time, or put it off until all product is delivered if you want. I've worked for small, privately held companies where I knew the accounting people, and I know that while changing the way we booked revenue or merging or splitting of a part of the organization often had innocent (i.e. valid business) explanations, the real explanation was often to obscure failures at the executive level. Failures to sell new product, failures to manage cost of delivery, or (and this is especially true of small companies) failures to manage executive perks which often sapped company resources.

In the 2000 time frame I worked for a large organization that signed a volume purchase agreement (VPA) with Microsoft. I don't know when Microsoft started using these VPAs, but I do remember reading that they changed the way they "booked" them about that time too although that didn't strike me as interesting at the time. The process was heralded by people higher up in the organization as a major cost savings, but as time went on and as activities associated with the VPA either happened, or in some cases were just stated to have happened, I began to suspect that there was no actual cost saving and that we might actually be spending more on Microsoft products than we would have otherwise.

I remember us having to come up with a count, which was really an estimate, of how many Windows PCs we (the entire organization) had. I know the count was more of a wild ass guess than it was a count, and the count was rounded up to satisfy Microsoft and (theoretically) get a lower unit cost. This organization was big, very big. Very few of our machines were on the Internet, and most of the machines were only networked locally, in many hundreds of LANs with local administrators. Had some new inventory process been imposed on these people I would have heard about it, and I didn't hear about it. Furthermore, most of these machines were not typical desktops, but were in place to run a specific set of locally written applications. They did not need to run Microsoft Office, for example, but the counting process glossed over that fact and even included the rights to run Microsoft products that weren't in use at all. The whole thing seemed to be a publicity stunt for certain organization managers, sounding great for the customer, but was really a windfall for Microsoft.

We were running Windows 2000, and nowhere near ready to convert to XP since in fact there were stragglers still running Windows NT, so in effect, we were paying a second time for software we already had. No worry though, said Microsoft, the contract includes an upgrade to XP, whenever we decided to do it, that is, as long as it is within five years, after which you do another VPA (I remember out VPA being for five years, but I don't think all VPAs have that term, some are shorter, I don't know if they come in longer terms as well). Nominally in fact the VPA was for Windows XP, it's just that other than writing a big check to Microsoft, nobody cared when or if the upgrade actually occurred. Do I have to spell out the rest of the story?

My guess is that the number of deals such as this is large (the entire Federal government agency by agency for a start) and when Microsoft makes claims about the number of 2000, or XP, or Vista licenses out there it's a lot of accounting tricks, after all, we didn't get an actual copy of any Microsoft products for each machine we ran. Instead we installed off the net, or from copies of copies of copies of the original disks. No need to mess with those fancy laser printed product keys. A single key made all the installs work without contacting the mother ship.

Yes, there was a costs "savings" for these VPAs, but the savings failed to take into account that facts that: (1) the machines were purchased with Windows already installed, (2) previous licenses had paid for the software again, (3) VPA1 had paid again, and (4) a subsequent VPA2 paid yet a fourth time. The savings MIGHT materialize if there were more frequent product releases from Microsoft (but guess who controls that) and only then if the customer were able to upgrade almost immediately (something the technical people know wasn't going to happen, but then the company/government negotiators are not generally technical people).

So, when Microsoft does their quarterly reports on how many licenses of various products they have sold I figure they are about as accurate as a weather forecast for this day next month. This is not just because I don't trust Microsoft, but because I think a lot of companies play these games. Maybe all of them do.

The difference between these "booked" numbers and what Microsoft deposits into the bank every quarter gives them a lot of room to paint a rosy picture when things are dropping off. If that is the case, and they make subsequent cuts or change their business in some drastic way those ruts in income stream can be smoothed out and the problem resolved without stockholders ever noticing. The more other activities the company is involved with, the more room there is to spread the blame around, making it look like the sacred cash cows are still in good shape while only these new (and ultimately expendable) ventures are holding things back.

Of course these book-cooking operations can't hide a monotonically decreasing income picture forever. The actions you have to take in the background get more and more drastic. Microsoft could use the billions they have in the bank to pay off any shortfalls they have, but that doesn't impress the stock market. If instead, you do something to drastically change the way you keep your books, say merge with another large company, spend most of your cash, stock swaps, redundancy layoffs... Some of these actions may actually improve your picture, but even if they don't you get an excellent opportunity to obscure the picture even further and a chance to promise shareholders that once the merger costs are absorbed, things will be wonderful again.

That's what I think is going on here, and because I think Yahoo has been playing similar games, no matter if the merger goes through or not, both companies are going to face dismal futures unless they make actual and substantive changes to their business models rather than superficial ones. (And did I mention the long term costs of ignoring your customers actual needs while you tinkered with your company spreadsheets?)

(This post revised and extended from an original Slashdot comment I made)