|
|
The February 1983 meeting looked back on not only COMDEX, but the second and third trade shows we attended: CADCON and CPM-83. The strains of converting a loose aggregation of entrepreneurially-oriented programmers into a responsive company can be seen in Dan Drake's announcement of the meeting. A large quantity of collateral communication led to the feeling that we had to resolve the issue of effective management.
In February of 1983, the company's cash position turned around and, in fact, doubled in one month. AutoCAD sales had begun to take off, and began a growth curve which, at this writing four and a half years later, has not yet seen its first inflection point.
The February monthly meeting will be at DUFF KURLAND'S meeting room (not Jack's) at 1:00 on February 6.
On February 13 we'll have the special meeting on managing the company. To approximate an all-day meeting, this is starting at 11:00 (eleven). It's currently scheduled for Jack's house. When originally proposed, way back in October, this was described as a brainstorming session; i.e., lots of ideas should be put forth before we start inhibiting ourselves by looking at them critically and realistically.
[Unauthorized editorial by DD: This is not a company owned and operated by some guy or guys in the Northbay. It is a cooperative venture of some entrepreneurs, all of whom think and act as if it were a company owned and operated by some guy or guys in the Northbay. This fallacy is self-sustaining and is very close to destroying the company.
[There are two equal and opposite errors concerning how to fix things.
One is to exhort everyone to buckle down, work hard, and act like an
entrepreneur. The other is to figure out better management strategies
for the people who run things to use in managing the workers. Can a
person who's miles out of the daily action, able to commit only a
small amount of time to the enterprise, really be blamed for not
feeling much like an entrepreneur? Will he feel better if there's a
better bureaucracy, more like the conventional company he works for
(considering that he's not so devoted to that company as to devote all
his time to it and resist new things like ADI![[Footnote]](i/footnote.png){kind=small}
Are there any
practical, concrete steps by which we can discipline or trick
ourselves to see the company correctly? Not that this will solve all
the management problems, but nothing else will without this.
[Finally, if you come to the meeting with a proposal that somebody should tie a bell on the cat's tail to warn us when it's coming, don't forget to consult your Boy Scout Handbook (to review knots) and your family attorney (to update your will). There's no use for proposals of the form, Let's You And Him Fix It This Way.]
Responsible opposing viewpoints will be given equal time; that is, if you have an idea you want circulated before the meeting, get it to me in writing (mail, MJK, courier, or whatever) by Tuesday Feb. 8, and I'll mail it to everyone on Wednesday. February General Meeting
by Dan Drake
The February general meeting was held on Sunday, February 6, at Duff Kurland's. Present were Dan Drake, Mike Ford, Dave Kalish, Duff Kurland, Greg Lutz, Mauri Laitinen, Keith Marcelius, Hal Royaltey, John Walker, and Roxie Walker.
There was no dissent from the minutes of the December and January meetings.
John Walker presented the financial report. As of January 30, one
year after the first organizational meeting, the company had a
positive cash flow, with net assets doubled from the previous
month.
Bank account: 7,966 Capital Preservation Fund 17,987 ------------------------------------------ Liquid assets 25,953 Receivables 7,734 ------------------------------------------ Total short-term assets 33,687 Liabilities & committed exp. (1,500) =========================================== Net 32,000+
Expenditures in 1982 amounted to $44,493, broken down as follows: hardware 48%; shows 22%; legal expenses 14%; printing 7%; stock repurchase 5%; others 4%.
The Board is now meeting informally during the week before each general meeting. At the meeting on February 2 the main topics discussed were marketing plans and manpower allocation.
The Board's idea, discussed and amplified at the general meeting, is that we should offer a standard distributor and OEM price of $425 or $450, and let them tell us what they can offer in return for the still lower price that they'll ask for. When people have new machines that they want us to put AutoCAD on, we'll try to negotiate an engineering fee for the work. If an OEM looks valuable, we may convert to his machine on speculation (as we have done up to now) and make an evaluation copy for him to look at, but without committing ourselves to anything in terms of an eventual release. Any manufacturer who is serious will give us a loan of his equipment for as long as we're supporting it in our software.
As to manpower, the Board noted that there would be no manpower for marketing AutoDesk if it were ready for release immediately; our marketing operation is overloaded trying to get full advantage out of AutoCAD. The implications of this were discussed later.
We are pursuing listings for AutoCAD in every relevant directory and are beginning to get announcements published. An article in ISO World got about 200 responses. We are sending announcements to about eighty industry analysts and a couple of hundred OEMs. An important goal now is to get a number of retail dealers around the country, who can answer questions and give demos. Of course, we went to two trade shows in January.
There
was a longish discussion of where we should spend money in
promoting the product, particularly whether we should buy into a
mailing of card decks by the S. Kline Newsletter. For instance, if
we burn most of the card decks we get, why think that everyone else
won't burn theirs? The upshot was that we should budget $1,000 for
Mike Ford to try out what he wants. It will take a month or two
before we have enough cash-flow information to allocate a proper
budget for advertising.
One of our objectives is not to be the Electric Pencil of the CAD industry (90% market share in word processing in 1978; zero in 1982). To do this we want to nail down as many manufacturers as possible, so that the next product to come along will find very little market left.
We are working with USI, which makes a cheap ($300) mouse and wants to use AutoCAD to push their product. We have sales now at Ford and Shell in England, with prospects for very large sales in the future. Houston Instrument plans to use AutoCAD to demonstrate its products at shows. Texas Instruments has an evaluation copy and is of course going to require 10,000 AutoCADs on its new IBM clone.
Dan Drake is now the interim sub-project manager for AutoCAD-86 (until
somebody complains about the cooking). A
large part of the job consists of talking to the outside world.
Technical people on the project (meaning Greg Lutz at the moment)
have unlisted phone numbers as far as the outside world is concerned,
and should be called as little as possible by our own people.
AutoCAD-86 release 1.1 went out around January 15 after a struggle with the document. There is a demo version, which is a full program except that it can't write any output on disk; it will also be fixed so that it can't plot from within the drawing editor. We solicit any suggestions for how people could get around the limitations to get useful work out of the demo version!
The next release is scheduled for March 10, with the full features of AutoCAD-80, plus the Epson screen dump on the Victor version. IBM Touchpen support, which is in the current version as an undocumented feature, will be officially supported in this release if we get a working Touchpen for IBM by March 1.
An evaluation version of AutoCAD-80 (and AutoCAD-86 on the IBM) is now running on the Vectrix, a beautiful $5,000 670×480 display with 511 colors selected out of a menu of 16,000,000. The version for the Aurora is released, and actually sold more copies than the Microangelo version in January.
The program now fills memory pretty thoroughly,
and the main
development project is to compact it a little more. It is also being
fixed to run on multi-user Turbodos systems.
There was some discussion of whether we should fix on PL/I (now that
the 8086 version is coming out) or on C as a single language for
AutoCAD. C, however, is impossible for AutoCAD-80, while PL/I has
much less future than C on the new machines, particularly the 68000.
This leaves us stuck with two versions forever.
Kern Sibbald is turning over the AutoDesk project to Mike Riddle. When Kern returns from vacation, he will start taking over the AutoCAD-80 project. AutoDesk is not close to a release yet, but as noted before, we couldn't effectively push it if it were.
There was no report, but the subject of conversion to the 8086 was discussed later in the meeting. The problem with conversion, now that CB-86 is supposed to be out, is the 3,000 lines of assembly code that have to be translated. The translation program for converting 8080 assembly code to the upward compatible 8086 is apparently worthless for large programs; therefore, the cost of conversion could be justified only if we had very good prospects for sales.
The only thing happening here is an attempt to get the books set up properly. This becomes critical as the date for income tax (75 days after the end of the fiscal year) approaches.
There was a post-mortem on the two trade shows (CADCON and CPM-83) in
January, together with COMDEX.
There was a general failure to plan for moving things out on the last
day of CPM-83. In the future our policy will be that at closing time
of any show we have the Walkermobile
and/or the Drakemobile
to haul things. Either of these can carry our signs and large chunks
of equipment, and is accompanied by someone who is supposed to know
what's happening.
There was a poll on the value of the three shows. COMDEX was considered very valuable and successful. CADCON was thoroughly marginal, with a low turnout of not very well informed people; and we didn't properly exploit the chance for a good look at the expensive competition. CPM-83 was much like a Computer Faire; the management of the show was much worse, but there seemed to be a higher concentration of dealers and fewer obnoxious people.
Lack of preparation and discussion before the shows sometimes caused people to feel they were making fools of themselves, as in trying to demonstrate what turned out to be an unimplemented feature. On the whole, we do much better than other exhibitors in giving out accurate information at our booth, but there's still room for improvement. It was suggested that everyone who will be working a show should take the whole of setup day off, so that we can do a tutorial on that day, either before or after setting up. There was no real consensus on whether we have reached the point of diminishing returns, where we're good enough at shows that there's no point in spending more effort that might be useful elsewhere.
We seem to need an AutoCAD Jockey. Like Computervision and other companies, we should find people who are especially good with AutoCAD and make them specialists in making the product look easy and impressive. Jockeys might be recruited from within the company or from end users. Even if they can't go around giving demonstrations, they can create drawings and set up demos for other people. This was agreed to be a good idea, but so far we lack anyone to take on the job.
An especially acute problem at shows is the taking of orders for products. At CPM-83 the mechanics of taking orders were haphazard—it was hard to find a clear space to write on or a place to file things—and we ended up with many undecipherable orders. The ideal solution would be to have one person doing nothing but taking orders; but it seems impossible to do that in a ten by ten booth.
There was agreement on some suggestions: All forms, manuals, VISA slips, etc., will be organized in neat, possibly color-coded boxes. If we have no other tutorial session before the show, there will at least be an indoctrination on the ordering procedures. There should be a special order form with “Take” and “Send” at the top to indicate whether the customer has taken delivery of what he bought. There should be a clipboard or something to guarantee a surface to fill out forms.
There was a time at CPM-83 when there was only one person at the booth. We must keep at least two people in the booth without fail, preferably three at a show full of thieves, like the Computer Faire (four is a crowd). This means that we should have four or five people at the show at any time.
As mentioned before, the Autodesk project is going to Mike Riddle.
Kern Sibbald will start taking over AutoCAD-80 when he returns from
vacation. Duff Kurland will start doing AutoCAD-86 work and will
investigate getting our documentation on a decent word processor,
which might be Perfect Writer.
There have been manpower shifts as described above. AutoCAD-86 has a manager (Dan Drake), who should get all phone calls; the people doing technical work are to be left alone.
The advertising budget is an ad hoc $1,000 this month, to be regarded as an expenditure of working capital. A real budget will be allocated in a month or two.
There are specific policies for improving our handling of shows, in terms of coverage of the booth and handling of sales.
The Solar System drawing was our flagship demo of the resolution of our floating point database and formed the centrepiece of many of the demos we did for venture capitalists in 1983. It was also the very first drawing I ever did on AutoCAD-86. I drew it in one afternoon on a Z-100 that I had just gotten to work with AutoCAD-86 (my first work on that version of the product).
|
|