we're model buyers, we should be model builders

A 'business model' we build. A personal, bespoke model that is implemented and sometimes refined. Accepted as dynamic. A matter of life and death for the firm.

Then we add the standard:

Organisational structures.
Accounting rules and models.
Software.

More models. Representation of reality. Generic and Generally Accepted square holes for any shaped peg. Rigid stuff.

In fact we put our lives (and the 'business model'!) into the hand of generic and bought models. Squeezed into them.

What if the models are wrong? Or even slightly wrong?
What if the models foster unwanted behaviour?

Then you lose, then the company lose, then society lose.

Business is organic - business is the people, with personalities, intuition, creativity and quirks. Thus no two business can be the same, nor should they.

Then per definition:

Rigid models cannot truly represent all of reality for all, thus the models will always be wrong to lesser or greater extent.

If you want to make better use of the resources, if you want a better place to work in, if you want to make great profits, in short, if you want to be successful - rethink the given models. Nah, do one better:

Become a model builder.

Organisational structures: See what you are, what you want to attain, where you want to go. Build a representation accordingly and use technology to strengthen your new model (blogs is a fine example if your wanted structure model is more inclusive and open than command).

Accounting: Sure, keep the GAAP for the authorities, but, blast it, should not the accounting model also represent how you want your processes to work and where you create value?
(If you like playing with numbers, here's one for you)

Software: Find software or create combination of software that enables you to build whatever you want, enable any process you can dream of, not something that copies what your competition does.
[DISCLAIMER: That's what I spend my time on, having fun, trying to build that kind of software stuff.]

Model building is fun and much easier than you think. Ask any kid.

Comments

[hugh macleod]

[cough] it's spelled "lose", not "loose".

[Ric]

The other problem with accepting existing models is that you are not differentiating your business from the others, so it becomes increasingly difficult to complete on anything other than price - a game in which there is only one winner. By having a greater diversity of business models, being different to other businesses, and seeking a global market to the greatest extent possible, there is an opportunity for more winners.

Having a 'global' view does not have to mean being physically global, but thinking like you are - don't narrow your goals to the domestic market - easier with digital products like Sig's software, but possible with physical goods like English Cut suits ...

[sig]

Hugh, and not only once, but three times! At least you could say I'm consistent :-)
Merci eh!

[Frank Patrick]

Sig -- Curious...Are you familiar with Goldratt's old circa 1980's attacks on cost accounting and the resultant development of "Throughput Accounting" as a bridge between "management accounting" to support rational operational decisions and "financial accounting" to support institutional reporting requirements associated with the tax man and wall street?

A lot of your language sounds consistent with those of us who have bought in to Goldratt's Theory of Constraints as a viable management theory.

[David Davies]

Sig, I found this through Frank Patrick posting to a Goldratt discussion group I frequent.

Are you aware of an English cybernetician called Stafford Beer (unfortunately now deceased)?

He developed a generic model of any Viable System. It is called VSM - Viable System Model and I think it addresses very well many of the topics you are interested in, particularly maintaining viability while also favouring autonomy.

As an advocate of Model Building versus Model Buying you should be aware of this body of work.

Best Regards and enjoyed your blog,David Davies

(PS. There is a commercial web site which has a lot of info at www.syncho.com)

[sig]

Frank, Goldratt and "Throughput Accounting" were completely new to me! But thanks to you and wonders like Google I just had a very quick update on his theories, and you're right, very interesting read it is. And quite parallel too in many ways, can see why you asked now!

In all humility I would say that I'd rather like to extend on what he says, using his basic idea of "old models" does not cover today's world: My tenet, as you saw in that other post, is that certain costs can be replaced with much cheaper means, information technology. Thus there is a (accounting) need to redefine many costs to be information costs. Like sales margins, distribution, marketing, inventories and much of what happens within the corporation proper.

I suspect my aim for that would be the same as for Goldratt, make current business more efficient by making the model that we steer by more true to current reality.
Being humble of course :-)

Thanks again, very enlightening that one!

[sig]

David, while posting reply to Frank you came in with even more info! Great! Google here I come, more to study tonight :-)

Much appreciated!

[sig]

Wow you guys, just have to add that after a dip-of-the-toes into the work of Goldratt and Beer - this certainly will lead to more than a quick study tonight! Excellent.

[MarkN]

One (extended) question I have is how you expect such highly personalized and stylized systems to be:

1) Integrated with the highly personalized and stylized systems of the:
- Company you just acquired
- The strategic partner you want to do business with
- The 5 companies you are joint-venturing with

and/or

2) Serviced in a cost effective manner by people who actually understand your highly personalized and stylized system.

In my view part of the benefit of the entrenched management methodologies and monolithic stacks of software that are out there is that there is a huge 3rd party market of complementary solutions as well as a large number of domain experts.

So what you really must be advocating is an uber disintermediation technology that somehow miraculously delivers unlimited customization and personalization whilst simultaneously guaranteeing universal compatibility and interoperability.

Forgive my skepticism.

[sig]

Mark, glad you asked - second to last paragraph: "miracously delivers unlimited customization"- Yes, precisely!

But miracles are sometimes closer than you'd expect: As you may have gleaned from other posts here, I'm an advocate of splitting information and logic in enterprise software architecture. With raw data only in the repository applying logic only when representing a model (in short delivering a report) then there are no limits to what you may deliver.

Interoperating with other systems is in principle just a cousin of the report - deliver a bespoke set of data to that other system, and voila, interoperability with no limits.

Where you should be sceptical is the other way round: If data is kept in a manipulated manner (as is standard for enterprise systems today), then the architecture I'm advocating would need the assumptions (in fact the whole logic behind the manipulation) as well as the data from the other system to make any sensible use of it. BTW, ditto for any non-standard use of those data today, see all the nice data mining and analysis tools being offered!

Good thing that my take on this is getting help from unexpected forces like Basel II and Sarbanes-Oxley and more. These new regulations in essence requires all data to be kept in original form. And I suspect those two are not the last ones we'll hear from!

[hugh macleod]

"In my view part of the benefit of the entrenched management methodologies and monolithic stacks of software that are out there is that there is a huge 3rd party market of complementary solutions as well as a large number of domain experts."

A short-term benefit at best, surely? Sig? Anybody?

[sig]

Hugh, if I'm being true to what I say I would have to say that it is only a question of time. How fast, how slow - I would not dare to guess.

But that the market is large is beyond doubt, and that it will be voluminous as in 'time x mass' as it peters out is also for sure. It is another example of the long tail, except that a shrinking market will make it more crowded unless some suppliers quit before time.

If that makes any sense :-)

[Ric]

Hugh: absolutely right - this 'large number of domain experts' will be equally 'experts' in whatever new flavour of software comes along (with just as much substantiation) - watch out for the sharks, Sig!

As for the complementary solutions, that is the whole deal with things like open source and stuff like Sig and Alan deal with, and the sort of things I'm interested in - small pieces of useful functionality with open interfaces that can be 'plugged' together to form potentially endless combinations of software usefulness (sorry - sound a bit like a marketer ...).

[Ric]

BTW Sig - I'm looking forward to a discussion - just need to get a working microphone for my Skype connection!

[MarkN]

Hugh, I personally hope it's a very short-term benefit. The reality is that if you are trying to raise venture capital for “new technology, old problem’ it is a question you must answer more precisely than "because I hope so".

Sig, the thing I don't accept in your thesis so far is that somehow there's a non-middleware solution to even the most basic issues such as database platform. Oracle? DB2? MSSQL? Postgres? MySQL? These are where 99.99% of the world’s data are stored. If the answer is that you'll abstract these platform dependencies away then guess what? You're selling middleware.

If the answer is that you'll create a new way of storing and representing data/information that is universally generic, then for now I would say to focus on one world changing, revolutionary, epochal thing at a time.

Let me close by saying that I have personally bet the farm on creating a domain specific microcosm of what you are proposing, so I assuredly agree with you conceptually.

Where I have a disconnect with how you are framing your idea is that as an expert software creator myself and long time observer of the enterprise and vertical business software industries I simply do not accept and indeed substantially discount "my software will revolutionize the world" rhetoric.

[sig]

Mark, love your points, and you sure know how to give a poor chap a big task here :-)

Let me do it backwards:

"my software will revolutionize the world": I would discount that too, and I hope I've never said it! I'd rather aim for "my software may be used to evolutionize the world". Up to the user. In other words, I accept the current situation, I accept that it may never change - thus I try to do a software that can mimic the current ways (*cough*.. as far as possible of course), and that can change towards whatever (not only what I think the future might look like) at the speed preferred. In the meantime I'll do my best to make the system able to compare favourably with the competition even while doing stuff the old-hat way. Kind of 'evolution as bonus' if you want. :-)
(No betting of farms please.)

"These are where 99.99% of the world’s data are stored": Well, I have no preferences. SQL, OODB, any make - where the data is stored or by what schema is really not important. What is important is in what form it is stored, raw or manipulated.
Exmple: Transactions (raw data) or end of month accounts (manipulated data). The former may be used for any accounting report possible (or even to build inventory reports, or more), the latter has to be translated and demanipulated to be useful for other applications.

"You're selling middleware.": What's in a word? Sure, I would not blush if that was the term. But for me 'middleware' is something that extracts data from a dataset produced by one application so it can be used as 'raw' data for another application which then applies another logic to the data.
What I would suggest is to store the information untouched and apply logic to it on the fly when you need a presentation/report. Then all 'applications' could use the same information without a need to descramble, i.e. the middleware concept would then be moot, or what?

"If the answer is that you'll create a new way of storing and representing data/information that is universally generic": No new way of storing, no new way of representing, no obscure schema. The crux: Actually no storage of representations of data! Only storing pure information so I can apply logic to it with any type of application or report generator.

Wow, that got long. Hope I made some sense... or at least that I was not completely senseless :-)

[sig]

Mark, just hit me that we've strayed a bit away from the post here and gotten rather deep into some aspects of the tech of what I try to build software-wise.

What I argued for in the post is that model building beats model buying (aka copying the processes of your competition) - and that is rather hard to attain with some of the larger enterprise software systems available. Hey one of them used "best business practices" as a slogan! In my warped brain equal to "copy your competition, be exactly like him!".

I even suggest that going there (building your own) is profitable! And that is not very revolutionary, is it?
Dell, Amazon, Southwest built their own models.

So if you asked me what I would hold out as my goal for what I build, then I would say "it's more like Lego, build your own castle!"
Then the middleware, database, structure chosen and whatnot is secondary (although important for the technical solution of course).

[NOTE TO SELF: Cannot shut up, can I?]

[MarkN]

Talking about something you love is always effusive.

As an aside, don't you think model building is anathema to transparency? Depends on how franca your lingua is, I suppose.

57 spreadsheets and no insight found (sung in my best Springsteen voice).

[sig]

Mark,
Hehe, don't I know it!

Heh, "model building is anathema to transparency?", now we talk...hmmm *finding Springsteen on iPod*...

A model is a "representation of reality", which in itself should have no bearing on transparency.
But, and this I suspect is where you are hitting the nail on the head: By use and perception a model may (often?) become a "replacement of reality", then more than transparency suffers!

Ahh *lightbulb*, that is what I've been chirping about: Models that even could be wrong, but at least they're models have been elevated to truth, to reality - and that has repercussions.

May I dare mention some models that have "replaced reality" and thus are defended vigorously as something "that has to be there"? Say organisational hierarchies, marketing, GAAP's as we know them. Did I mention middleware? No, joking, just joking! :-)

What do you think?

[Rockster]

Mr. Org Theory here again... 3 points tonight...

1) People are generally lazy... (from a recent Seth Godin post, "What Every Good Marketer Knows"):

People are selfish, lazy, uninformed and impatient. Start with that and you’ll be pleasantly surprised by what you find.)

Therefore, people REALLY LIKE pre-packaged stuff.

2) "Organizations" exist as "real things" in large part because they are the evolved "survivor routines" of sets of routines that have been tried and shown to work for the particular set of problems the organization faces. Without routine activity, every day at work would be starting from scratch... Even highly creative enterprises, consultancies, etc etc that supposedly "solve unique problems" end up with a playbook.

3) The reality, therefore, is that organizations will tend to die before they change to meet a new reality. And that's a good thing.

So here's what I'd add: when you build your model, it is a unique adaptation (your theory, actually, of how the world's working right now). Therefore: KNOW and BUILD IN the process to kill the model before it gets really messy. (won't happen... people are generally lazy and fear failing, so they'd rather not know).