Abstract

On data processing-related matters one has to buy adequate results, not work or matter.

Customers need solutions

Please read this old Apple ad from the early '80s:

(macmothership.com publishes many Apple ads)

The mention If you're looking for a personal computer which solves problems instead of creating them is very close to a good'ole idea of mine: an adequate tool is a friend, not a foe. As wrote St Jérôme (comment. in Ecclesiastem): pereant qui ante nos nostra dixerunt (translation: a pox on those who have declared our bright ideas before us).

This ad conveys a message: the computer is a tool but the customer searches a solution, not a tool or a service.

Solution means here a fix to a problem or an enhancement at the functional (not technical) level, guaranteed by a single supplier.

Important matters are important. Do your homework!

Letting an employee design, elaborate, deploy then maintain all solutions leads to suboptimal results because no one can master all fields (Web, office automation, specific (vertical) software development..). Dedicating a bunch of employees to a given family of solutions seems only sound in a big corporation or if those solutions are of paramount because some part of the business depends on it. Better cope yourself with any critical process.

Therefore all corporations have to buy adequate solutions, at the very least for non-critical processes. Let's devise an adequate approach.

The all too common 'hit and run' seller strategy

Right now a vendor sells a:

thingie, then often runs away
service, then often gets paid as long as he stays, therefore, in order to stay around and keep invoicing, he tries hard to find "problems" and potential "enhancements"

The market needs efficient (therefore coherent) behaviors

There was a classic way (was it in China?) to sell some 'medical service plan': the physician is paid as long as the customer stays healthy, copes with a limited amount of patients and has to observe some latency before replacing a dead one. This is adequate because it does not distort the behaviors: the customers knows what he buys ("a better health condition!") and the seller always fights to provide it, by solving problems and enhancing the global condition, because he has nothing to gain from not dedicating himself to delivering the most useful results.

This approach implies a common (shared) definition of 'good health', which is useful because in such a deal everyone needs to understand the context, the constraints and the objectives. This very definition leads to a protocol able to show, at any moment, if the 'solution' works as required. The protocol, which is in fact a delivery-test (reception-test), is derived from the definition and established before the deal. It expresses the customer needs, with as few technical mumbo-jumbo as possible, by specifying and testing the resources needed along with the inputs and outputs.

The IT world may benefit from this approach. It somewhat started when people thinking I don't want to become my supplier's hostage began to use open source software.

Pertinent conventions

Therefore a solution-approach may be the most adequate in the IT world, for all concerned. It implies:

the 'solution' is defined, a set of tests is devised, then at any time and for all parties if the tests run OK, then the solution works

there is a neutral (non intrusive) and fast way to test at no cost the solution, which provides concise results (this part works, this other one does not) accepted by all parties. All parties can use those tests at will. Moreover provisions are made in order to quickly extend such a definition.

all parties thrive in order to maintain the solution up and running

the customer periodically pays, as long as the solution works. After a malfunction, the service company pays penalties proportional to its relative importance (each test is associated to a penalty) and duration. Note: all tests will probably be implemented in 'defensive' mode, meaning that they will first check all needed resources. This is in line with the efficient 'test-driven' development.

the customer is not a hostage

he can abandon a service company, therefore:

for each need there are many competing solutions
all customers demands interoperability, especially open standards
from time to time the tests are redefined, letting any party adapt or break the deal

Transition towards this market state

If big customers demand this 'solution' approach, the first service company to comply will apparently reduce its benefit because it will not be able, as usual, to transmute its customers into hostages and to create problems in order to make them pay. This service company will instead keep them, not against their will but simply because the service/cost ratio will be judged adequate (with respect to market indicators). Then new customers may be attracted. The service company will probably have to fight against some traps (probably massive disinformation) set by its competitors using the classic model. But I bet this will prove to be an efficient strategy.

Ack

Thanks to WJ.

Other texts