Start in the Middle

One of the great puzzles of software design is how to create a user experience that is both intuitive and enjoyable while enabling the user to get the job done with a minimum amount of hassles.

Too often software (or web browser-based services, for that matter) is designed to accomodate all forseeable use cases, which leads to factoring a bunch of edge cases front and center into the design.  This results in software that is cluttered, confusing and inefficient. 

Why?  Because rather than enabling the user to focus on the 20% of the software functions that they will use 80% of the time, the user is perpetually surrounded by a bunch of noise that, while novel and impressive on first run, places a perpetual tax on the user across the lifecycle of using the software. 

The goal of my post is not so much to try and analyze good and bad approaches to software usability and user interaction design (a great book on this topic is, "The Inmates are Running the Asylum") as it is to offer up a method that I have found pretty useful in working through complex UI and workflow design challenges.

I call it 'Starting in the Middle,' and the idea is that when you encounter a UI design challenge, you begin not with the beginning of the workflow that is to be facilitated with your software, but start in the middle of a particularly important task. 

The reasoning is two-fold.  One is that by starting in the middle of an important workflow you naturally focus on the specific job that your user is going to want to accomplish at this place in the software. 

This ensures that you focus on the substance of the job -- i.e., named procedures, logical ordering of core functions, natural leverage or entry points from other places within the software and subsequent handoffs once the task is completed.  In other words, you focus on substantive outcomes, internally logical and externally synchronous actions and avoid getting distracted by global constraints and style points.  For me, it turns the process of UI and workflow design into a kind of orchestrated chess board move, which makes it surprisingly manageable.

The other reason for this approach is that while all tasks are ultimately bounded by the user interface and workflow constructs that you have built into the software, invariably a few different workflow types may be supported within your software.  In my experience, one size and one design does not necessarily fit all when talking about meaningful functionality with user lifecycles. 

Hence, in the process of revving the software or adding on new functionality, you may conclude that different function sets (or more likely, complete tasks) fit more logically within one workflow type than another.  Or, you may conclude that the new functionality sufficiently re-defines the job you are enabling a user to accomplish that significantly iterating existing workflows is warranted.

I should disclaim at this point that I am not an engineer and I am not a trained user interaction designer.  I am just an entrepreneur who believes that the marriage between product design and product marketing is essential. 

This is so because the jobs your target users ultimately will hire your software to perform for them will need to be aligned both inside the product and outside it (when formulating your go-to-market strategy, messaging and the like). 

I have also found myself in more than a few startups where I was the guy raising his hand to take a swag at the user experience side of the software.  (Of course, this is/was always done in tandem with a great developer or two who has to implement the functionality and in the process becomes a key collaborator on what ultimately gets baked.)

In any event, next time you face a difficult user interface or related workflow challenge, don't start at the beginning.  Start in the middle. 

Comments

[John "Z-Bo" Zabroski]

I agree with your principles and follow your analysis but user experience is mainly determined by the users, their locus of attention and the automaticity of tasks. The more users, the harder it is to control and measure these three factors. I've always believed the best design strategy is to let the users regulate these behaviors as much as possible.

Every tweak the designers make the product should be to amplify marketability. Marketability and usability might seem conjoined, but rather they are complementary peices. A usable product can be marketable, but without facilitating a marketing functionality in a hassle-free manner, the product will not be enjoyable to use.

Just count the number of marketing mechanisms on a number of websites, and rate their effectiveness from one to 10. These are the coefficients that determine how likely people will use the site compared to another site with the same exact content. This is a really simple scorecard technique I made up to evaluate the marketability of a website and I like how it performs (it is a mathematically tractable linear programming model).

[Mark Sigal]

I don't disagree with any of what you are saying John, other than that I would net out what you are saying along lines of earlier post I wrote on jobs, outcomes and constraints.

Marketability ties to specific jobs, outcomes and constraints, and you build your software asking first who is the audience and what are they hiring for, for what outcome and what constraints do they face.

Over time, as you support more types of jobs, you need to come up with a holistic model that factors the different modes users will be working in and minimizes the noise from adjunct modes when you are not in a given mode.

Where lots of software goes down the crap hole is that the design is built around a "composite user" which attempts to aggregate disparate individuals and aspirations into a unified whole, which usually pleases no one. Microsoft products are the ultimate example of this approach.

A final point that is that designers/developers/marketers assume that customers can just tell you what they want or that you can just watch what they do and figure it design from there.

The truth of the matter is that there are different orders of problems being solved. In some cases, you know what you need to know and it is a simple matter of plugging it in. A protocol would be a simple example. In another case, you don't know what you need to know but can readily get it. How to make lasagna is a simple example. In other cases, you don't know what you need to know and the answer is not readily available, which requires a fundamentally different approach to solving the problem. Cold fusion is an extreme example.

Similarly in building new products and launching new markets you need to know how well formed the problem is, how uniformly expressed the solution is and go from there.

When Ford built the Model T he couldn't exactly ask users what they expected out of an affordable sedan. :-)

Mark

[John "Z-Bo" Zabroski]

Designing a product or service to be modal is a flaw in my eyes.

Modes mean conditions. I don't believe in conditions. I believe in fewer conditions. I want to use a product unconditionally -- the way I feel like using it.

You are right that Microsoft tends to suffer from building modes into products. Microsoft would be more aptly named Monolith.

I am not sure what you are referring to with your cold fusion comments. There's many types of "cold fusion," some of which work (muon-catalyzed fusion) and some of which don't (Fleischmann-Pons). I don't see a difference between cold fusion and hot fusion research. All atomic energy studies focus on getting higher net energies (energy input minus energy output). The only layman's difference between cold fusion and hot fusion is cold fusion would mean less nuclear radiation (safer, more stable energy).

I follow what you are saying in terms of launching new markets, but I think what you are trying to tell me is something you have a feel for and I do not (yet). How to tackle a new market is probably something you are well-versed in (or at least getting good training in right now).

[Mark Sigal]

Jobs, by definition, are modes of operation. There is big difference between trying to be all things to all people (Microsoft products), trying to find one model that works for all (the lowest common denominator) approach, and working for the perspective of facilitating specific outcomes and workflows.

TypePad, for example supports different supports different logical jobs -- creating the look and feel of the blog; create a post; administer the blog; extending the blog with blogrolls, video rolls and other widgets. The beauty of what they built is that it neither shoeshorns a bunch of functionality into one bucket nor tries to make the blog ownership lifecycle a one size fits all schema.

[Herakliusz]

What I want to get is a edited, formated content. With value added. Whether be it made by users or media. I want to be guided and redirected in my search for the next content. The machine exists only for this purpose. So much for usability isues. Those two areas - content and usability - represent my own, user's emotions that I want to express, even by passively watching a page. I'll remember where did it happen.

[Mark Sigal]

Thanks for the comments, Herakliusz. Usability is absolutely critical to what I am talking about, as is context and the workflow that ties it together.

[WLADIMIR GUGLINSKI]

I would like to say some words concerning what said by the journalist Bob Weber:
"Regardless of experimental results, one needs a convincing theory of CF"
in the link:
http://www.strategykinetics.com/2006/02/cold_fusion.html

Before to understand cold fusion, we neeed to have a complete understanding of the nuclear phenomena. However we dont have it.

In the Introduction of my book QUANTUM RING THEORY, it is written in the page 4:
....................................................
“Perhaps one would like to say that the foundations for cold fusion are the same of that proposed in Quantum Mechanics. Indeed, in Jan-2004 the cold fusion researcher Dr. Dimitriy Afonichev sent me an e-mail where he said the following:
‘I think that occurrence of cold fusion can be explained on the basis of the existing theories’.
Truthfully his words transmit not merely a personal opinion, because actually several theorists those try to explain the cold fusion occurrence share his viewpoint. However such opinion is very intriguing, since the own academic community is agreeing that the existing theories in the branch of Nuclear Physics are unable to explain even the ordinary nuclear properties, as confessed by Eisberg and Resnick in their book Quantum Physics, where they say in the first page of the Chapter 15:
‘Though we dispose nowadays of a sufficient complete assembly of information about the nuclear forces, we realize that they are too much complexes, not having been possible up to now to use this acknowledge for building an extensive theory of the nuclei. In other words, we cannot explain the whole properties of nuclei in function of the properties of the nuclear forces that actuate on their protons and neutrons’.
So, as the existing theories are unable to explain the nuclear properties responsible for the hot fusion occurrence (which occurs according to the principles of Quantum Mechanics), it's hard to believe that such existing theories could explain nuclear properties that would be responsible for the occurrence of some so much complex as it is the cold fusion (which occurs by infringing the principles of QM). “
....................................................

For a layman to understand easily that said in the Introduction of my book, take for instance the interaction between two neutrons.
Two neutrons have no repulsion. But in a short distance, they are attracted by the strong force. So, after interacting within a nucleus, two neutrons would have to form the 0n2, and would never separate anymore.
But 0n2 does not exist in nature. Heisenberg tried to explain it with the introduciton of the concept of Isospin. Unfortunatelly the isospin is an abstract mathematical concept.
Two neutrons tied strongly by the strong force cannot be separated by an abstract concept, because an abstract concept cannot produce a FORCE capable to win the force of attraction by the strong force.
Only a FORCE of repulsion can win the force of attraction.
A NEW NUCLEAR MODEL (that shows what is the force of repulsion between two neutrons in short distances) is proposed in my book Quantum Ring theory.

In 2002 the Infinite Energy magazine has published my paper “What is Missing in Les Case’s Catalytc Fusion” , in which I have proposed some improvements to be addopted, in order to avoid the missing of replicability.

In 2003 in the ICCF-10 Lets and Cravens exhibited their experiment, in which they have adopted the suggestions of mine in my paper published in 2002 by IE.

In my book I propose an explanation for Lets-Cravens experiment, showed in paper entitled “Lets-Cravens Experiment and the Accordion-Effect”

The Accordion-Effect is a nuclear property unknown by nuclear theorists, and it is responsible for the resonance that takes place between a nucleus (for instance Pd) and the oscillation of deuterons due to zero-point energy.

After reading some of my papers, the late Dr. Eugene Mallove said in 2004: "Guglinski has interesting and intriguing ideas".
That's why he suggested to put my papers on a book form, and to publish it.

However, Dr. Mallove did not read my papers concerning the new nuclear model.
.
.
.
.
.
=======WHY COLD FUSION IS NEGLECTED BY ACADEMICIANS=======
The stronger reason why the scientific community neglects cold fusion is because its occurrence requires a neutron model n=p+e formed by proton and electron. However such theoretical model violates the Fermi-Diract statistics.

A model of neutron n=p+e that does not violate Fermi-Diract statistics is proposed in the book QUANTUM RING THEORY (QRT).

Two papers on the neutron new model n=p+e of QRT are available in the Internet.
They are:

1) ANOMALOUS MASS OF THE NEUTRON
2) NEW MODEL OF NEUTRON

Before to post here the two links, I would like to give some enlightenment on the paper NEW MODEL OF NEUTRON, as follows:
.
.
.
1) When we analyze the mass of pions according to the current Standard Model, we arrive to contradictory conclusions about the mass M(d) of the quark down and the mass M(u) of the quark up.
In the paper NEW MODEL OF NEUTRON it is shown that we arrive to the following two conclusions:
CONCLUSION 1: M(d) > M(u)
CONCLUSION 2: M(u) > M(d)
.
.
.
2) Look at the chemical reaction Na+Cl->NaCl
QUESTION: what is the matematical formalism underlying such a chemical reaction?
ANSWER: No one. The chemical reactions have not been established through the mathematical formalism.

The chemical reactions have been established based on the LOGIC, and such a procedure was viable because the chemists had the help of a property of the chemical reactions: the mass of the reagent elements does not change after the reactions. For instance, the mass of Na is the same in the two sides of the equation Na+Cl->NaCl.

In the case of the high energy nuclear reactions the discovery of the equations became very complicated, for two reasons:

1) Either particles can desintegrate by discharging energy, or particles can be created, by the transformation of energy to matter.

2) In the model adopted by the theorists, the addition of spins is applied to all the reactons.
However in the beta decay the addtion of spins cannot be applied (but there is conservation of the total angular momentun, because in the reactions there is creation of neutrinos and antineutrinos).

Such anomaly in the addition of spins in the beta decay made the situation to be very bad, and the theorists could not apply the LOGIC for the discovering of the mechanic of high energy reactions, as the chemists made in the Chemistry.

That’s why the theorists tried to solve the problems by the mathematical formalism, through the Lie symetries as SU(2), SU(3), etc.
But the result was unsatisfactory, as one can understand easily. There are particles that does not fit to the theory, and that’s why Murray Gell-Mann felt the need of proposing ad hoc bandages, like the Strangeness.

As the theorists did not discover the true cause of the beta decay anomaly, they impute to other cause the occurrency of that anomaly: they state that the parity is not kept in the beta decay.

By addopting the “spin-fusion” hypothesis proposed in QUANTUM RING THEORY, it is explained the anomaly of the beta decay, and from such a way the high energy reactions can be explained through the LOGIC, in the same way as occurred in Chemistry for the establishment of the chemical reactions.

The two links are:

NEW MODEL OF THE NEUTRON:
http://www.geocities.com/ciencia2mil/NewMODELneutron.html

ANOMALOUS MASS OF THE NEUTRON:
http://www.geocities.com/ciencia2mil/NEUTRONmodel.html