By Darrell Mann
The algorithm for inventive problem solving (ARIZ) schema – in all of its many forms – is sometimes considered to be ineffective or too-slow by users. The paper explores the underlying reasons why such problems occur by looking at how these types of processes match up the workings of the human brain.
Consider a simple case study involving the re-design of a clothes-peg by reading the following summarized transcript of what two people involved in the project did:1 (The numbers in parentheses refer to information shown in Figure 1.)
Person A: Let's re-design the clothes-peg. (1)
Person B: To do what?
Person A: Make it simpler to hang out washing.
Person B: What's wrong with the current designs?
Person A: Well, I was thinking about it. One thing is they require two hands. One to hold the clothes – one to hold the peg.
Person B: So you think we need a one-handed pegging system? (2)
Person A: Something like that.
Person B: That means that one hand must be able to hold the peg and the clothes. At the same time.
Person A: Maybe. Sounds like a contradiction if that's the problem.
Person B: We want to increase ease of use, but there are too many things to hold?
Person A: And they could all be different sizes. Socks. Sheets. Etc. Area.
Person B: Matrix?
Person A: Matrix. Here we go. 35, 1, 13, 17, 14. 4, 12, 2. (3)
Person B: We need to get something around the clothes-line to grip. Holding it in the same hand as the clothes.
Person A: Needs to stick to the hand?
Person B: Curls around thumb? Principle 14. (4)
Person A: Sticks by itself? (5)
Person B: Or. We should think about IFR. (6) No peg. The clothes stick by themselves?
Person A: The line sticks by itself. (7)
Person B: Self-fastening clothes-line. (8)
Person A: We need to get the clothes off again though. (9)
Person B: The next contradiction already.
Person A: We want the line to stick and not stick. (10)
Person B: Separate on condition? Stick when clothes are wet, then release when dry?
Person A: And fall on the ground.
Person B: Maybe not. Finger temperature?
Person A: Shape-memory! (11)
Person B: Cold day though. Cold hands. (12)
Person A: You won't put the clothes out on a cold day. (13)
Person B: Hmm.
Person A: So what then? Post-It? Glue? Re-usable? (14)
Person B: We should go and have a look at the function database. "Join." Or "un-join." "Separate."
Person A: Join. Lots of possibilities. Vacuum? (15)
Person B: Too complicated. (16)
Person A: Magnets? (17)
Person B: Magnetic line? (18)
Person A: And magnetic peg. That might solve the one-handed problem. (19)
Person B: Maybe. Heavy though? Another conflict. (20)
Person A: How much force do you need? (21)
Person B: Let's look at it later. Come back to the Database. What other ways? (22)
Person A: Static? (23)
Person B: Nope. (24)
Person A: Here's one. Gecko! (25)
Person B: Gecko?
Person A: Gecko foot. Sticks to anything. Then comes off again. Walking across ceilings.
Person B: Mmm.
Person A: It's brilliant. Gecko-line!
Person B: Sounds too good. It'll damage the clothes. Not enough grip. (26)
Person A: No it won't.
Person B: How do you know that? What do you know about Gecko feet?
Person A: I think it'll work.
Person B: Hmm.
Person A: And even if it doesn't, we just have to solve the contradiction. (27)
Person B: Hmm. I like the magnet better. (28)
Person A: Force versus something. Reliability? (29)
Person B: Durability?
Person A: Flexibility?
Person B: I still prefer the magnet idea. (30)
This is fairly typical of what happens at the "fuzzy front-end" of an inventive problem situation. It is necessary for problem solvers to, as architects describe it, "find their way in" to the situation. During this activity, there appears to be considerable "to"-ing and "fro"-ing between problem defining and solution generating modes. Figure 1 illustrates the sequence of events through the lenses of such shifts, including divisions of both definition and solution tasks into "generic" and "specific" regions. The two people oscillate frequently among all four levels of thinking.
The numbers in Figure 1 correspond to the numbers in the earlier transcription.
Assuming this kind of scenario is realistic – and "wicked problem" expert Jeffrey Conklin would say it was – then it has something fairly important to say about attempts to construct step-wise and sequential problem-solving systems.2 Perhaps most importantly, it says that matching up these steps and problem-solving systems will not work.
The evidence that it does not work is everywhere. Perhaps it is most apparent in the software domain; the phenomenon of customers not knowing what they wanted until they saw the first prototype, and then realizing they wanted something extra – or something completely different – is the bane of the software engineer's life. But this may be a fundamental part of the way the brain works – a person needs to see something before knowing if it was wanted.
The brain is a wonderful instrument, but it seems to possess a number of characteristics that fight hard against attempts to impose fixed ways of doing things. For one thing, most brains seem to be much happier during solution generation than in problem definition. Once a clue about solving a problem appears, in other words, the human brain is magnetically drawn toward exploring that solution even if it not time to be in the solution generation mode.
What is the blinding flash of the obvious that designers of sequential problem solving processes have somehow failed to address before? Step-wise processes will not and cannot work so do not try to create them.
In truth, many have long been conscious of the oscillatory nature of working on inventive problems. The tide of other methods works against problem solvers, for seemingly every other problem solving scheme has fallen into the same hole of psychological inertia. The hole says there must be a process and that all processes must follow a one-two-three-four sequence. The worst offender of them all is, perhaps, ARIZ. From its initial inception through all of the myriad versions to software tools, there has been the implicit assumption that problem solving procedures must be step-wise and sequential – step "n" should not be carried out until step "n-1" has been successfully completed. This seems to be a key reason why TRIZ – and every other creative problem solving technique – has failed to reach a critical mass of acceptance.
If this is in any way true, there is a contradiction: how can people work to create a linear, reproducible, structured process when the brain works in a non-linear, unstructured way? How can problem solvers be simultaneously systematic and non-sequential/structured and un-structured?
One option is to solve a problem in any sequence the user desires, rather than following the defined structure. But this option does not give the user permission to jump out of problem-definition mode thinking into solution-generation mode. This may happen unofficially, but this has never been a defined option in a problem-solving process.
The whole point of the structured problem-definition scheme then becomes the desire to put all of the available definition strategies together (into a linear sequence, if so desired). But not just to put the strategies together, but together in such a way that, whenever a jump to solution generation occurs, it is clear where the people are in the definition process in order to return to the place they left and work through the other remaining definition tools.
The whole point of the structured problem-definition scheme then becomes the desire to put all of the available definition strategies together (into a linear sequence, if so desired). But not just to put the strategies together, but together in such a way that, whenever a jump to solution generation occurs, it is clear where the people are in the definition process in order to return to the place they left and work through the other remaining definition tools. Figure 2 illustrates this basic scheme.
Each white rectangle represents a systematic
This scheme has proven to be less frustrating for problem solvers during problem solving sessions. In one particularly striking case, when trying to answer the first "where are you trying to get to?" a conflict between what management wanted and what the customer wanted was identified. That conflict was transferred that problem to a different box (tool) and produced a breakthrough solution within 10 minutes.
The next challenge is to encourage problem solvers to select the templates in a random or "when you feel ready for it" manner. There is never an "optimum" (the enemy word of innovation) sequence, but arranging by degree of difficulty appears to be a good option – particularly for newcomers.
Some recent case studies further demonstrate the sequence of tools and strategies problem solvers have employed. One case study involves full-time TRIZ/systematic innovation users; the second case study involves relatively new users. The latter group is primarily from the education programs run by the Hong Kong government and from delegates enrolled in innovation certification programs.3,4 In the second group, delegates were typically shown the default sequence of tools and templates and given the instruction "if you find you are not benefiting from a particular tool or template, or wish to try one you like better, please feel free to do this (remember where you were in the sequence though, so that you know where to come back to if you get stuck)." No instructions on sequence were given to the experienced users – the sequence of tools used was recorded retrospectively.
In all cases the full array of available tools and templates were as follows:
(Note: D1-D10 list follows the same sequence as the "define pack" found in reference 5)
D1: Initial definition – where are we trying to get to?
D2: Problem re-definition ("why-what's-stopping" analysis)
D3: Function and attribute analysis
D4: Perception mapping
D5: Resources (including evolution potential assessment)
D7: Sore point identification (root contradictions)
D8: Ideal final attributes
D9: Ideal final result (IFR)
D10: Maturity analysis
D11: Root cause analysis
D12: QFD house of quality
D13: Other (e.g., Kepner-Tregoe, axiomatic design and other tools already used by the problem solver)
(Note: Tools S1-S9 are TRIZ-based)
S1: Contradiction matrix and inventive principles
S2: Trends of evolution
S3: S-fields and standard solutions
S4: Effects/knowledge databases
S5 : Resource exploitation (combination, sub-division, other strategies)
S7: Smart little people
S9: Omega life views
S10: Random word (Edward DeBono)
S11: Direct analogy
S12: Feature transfer
|Table 1: Tool/Template Sequence Used by Problem Solvers on Different Case Studies|
All of the cases terminated at a point at which the problem solvers deemed that the solutions generated met the desired outcomes. The main message that Table 1 demonstrates is that both relatively new and experienced users all deviated to at least some extent, and many considerably from the default sequence of tools. In the case of the relatively new users this happened despite the fact that they had not been presented with any sequence other than the default sequence. Other general observations include:
The human brain has evolved to remember patterns in order to make predictions about what will happen next.6 This is seen anytime a person is presented with a new problem to solve. The first thing that person asks is "have I seen this problem before?" If the problem is familiar, the person retrieves the already existing "script" that tells the individual how the problem was solved. If the problem is new, the brain looks for the closest thing to the problem. Here the hierarchical nature of the brain's storage structure comes into play.7 If a complete script for the problem is not available, the brain looks for elements of the problem (scriptlets) that match.7
This mechanism is important in terms of problem solving because a "this is what we do when we solve a problem" script exists for every problem solved in the past. Inevitably, therefore, there are numerous existing scripts and scriptlets for problem solving. Brain tend to take a path of least resistance and so because these scripts and scriptlets exist, the brain is highly inclined to use them rather than try to re-write – or build new – scripts.
The desire to switch from problem definition to solution generation mode is one of these scripts; the brain does not like the unexplained and unexpected and rapidly tries to resolve discrepancies between what was expected and what actually happened (expectation failures). When presented with a new structured problem solving process like ARIZ – a process that almost never matches an existing script or scriptlet – it feels unnatural and, worse, the brain is more inclined to either try and fit the new problem (itself, by definition if it is an inventive problem, one that cannot have an existing solution script) to an existing script or even ignore it and simply continue to use the existing script. The net result in both cases is that a repetition of the the clothes-peg problem solving conversation summarized earlier.
Some TRIZ practitioners talk about the "magic three" – work through the tools and processes three times and a person will begin to build a TRIZ script (or, more likely, smaller subset scriptlets like "use the contradiction matrix"). For the early sessions, the person is in a state of conscious competence – consciously working through the script – and likely experiencing expectation failures. What a TRIZ tool was expected to do, turned out wrong. These expectation failures are critical; these are the times when the learning takes place.7
The "path of least resistance" response to these failures is to give up and revert to already existing scripts. What should happen in the presence of failures is to plant the idea that the failures are as important as the successes. Failures need to be explored until the uncomfortable failure feeling can morph into a positive resource. This may be the trick to not only getting to the magic three state (evidence suggests that 50 percent of people will take the path of least resistance and not reach this point), but also to the unconscious competence stage.
Unconscious competence is the state at which the expert users (featured in Table 1) operated. These are TRIZ users who have built well-established TRIZ scripts and scriptlets. Table 1 illustrates that the sequence of tools used on different problems is markedly different. What may be happening here is that these unconscious competence users are now unconsciously making connections between what they see and hear in the description of a new problem and the already existing scripts. One of these scripts is the important one – the one that says that linear, sequential processes do not work. Processes that are simultaneously structured and unstructured are needed for problem solving.
Darrell Mann is an engineer by background, having spent 15 years working at Rolls-Royce in various long-term R&D related positions, and ultimately becoming responsible for the company's long-term future engine strategy. He left the company in 1996 to help set up a high technology company before entering a program of systematic innovation and creativity research at the University of Bath. He first started using TRIZ in 1992, and by the time he left Rolls-Royce had generated over a dozen patents and patent applications. In 1998 he started teaching TRIZ and related methods to both technical and business audiences, and to date has given courses to more than 3,000 delegates across a broad spectrum of industries and disciplines. He continues to actively use, teach and research systematic innovation techniques and is author of the best selling book series Hands-On Systematic Innovation. Contact Darrell Mann at darrell.mann (at) systematic-innovation.com or visit http://www.systematic-innovation.com.