Innovating from a Clean Slate

A common approach to improving a product – whether a physical good, a service process, or some combination of the two – is to begin with the product as it currently is, identify problems that create customer dissatisfaction, and solve those problems.   It is an entirely sensible practice, and one that is capable of making many small, incremental improvements to a product – but it utterly annihilates any possibility of making significant and revolutionary improvements that will create innovative improvements that amaze customers and leave competitors far behind.

The reasons for this are straightforward:

• The as-is/to-be practice makes the assumption that the “as is” state is generally acceptable and needs only minor adjustments, so nothing revolutionary is ever considered
• The same process tends to identify many small problems that are easy to address, so people spring to action to make “quick wins” while the bigger and stickier problems are ignored or unaddressed.
• Where physical products and service procedures are already similar across firms in a given product category, all firms that use the as-is/to-be comparisons see the same problems and derive the same solutions, ensuring that products remain commoditized

Revolutionary improvements require revolutionary thinking, which is a different process to examining and making minor adjustments to what is known.  Innovation on this level requires clean-slate thinking, going back to the very basic assumptions about the problems that customers are facing and the ways in which products provide solutions.

As an example, consider the process of learning a foreign language.   The traditional approach to teaching language is incrementally building vocabulary and syntax: the student learns a number of words and follows models for how they can be arranged into sentences.   This model has been used in both academic courses and professional training, and is highly ineffective: it’s hard to learn a foreign language because it is not taught very well.

For many years, there was not much improvement in teaching methods, because education providers focused on improving the parts of their pedagogy without reconsidering the entire system of teaching.   That is, those who sought to improve instruction came up with different sets of vocabulary words, tailoring courses to business discussions or the common problems of international travellers – but learning vocabulary was still a matter of rote memorization of the names of things and actions out of context.  Likewise, it was recognized that classroom education (having to be in a certain place in a certain time) was inconvenient, so lessons were recorded into books, tapes, and videos that could be portable and consumed at the learner’s convenience – but the method of teaching was still the same.

These incremental improvements addressed specific inconveniences of the learning process, but kept the process the same – and the process, itself, remained broken.  As such, learning a foreign language remained a difficult and time-consuming process, and one which most people who had a desire to learn another language chose to avoid entirely, giving up on achieving their goals.

Solving this problem, and revolutionizing the language-learning problem, required stepping back from existing practices to ask the question, “how do people best learn to speak a language?” and to consider, simply enough, the manner in which a person who moves to a foreign country learn the language without classroom instruction.

Considered in that manner, it becomes obvious that people do not memorize lists of words in isolation, but learn words that they hear in the context of everyday life.  This goes both for vocabulary (how a given thing is called) as well as syntax (how an action is described in a way that identifies who is performing it and when).   The core problem, which the system ignored, was the manner of learning, not the content of the lessons.

If I’m not mistaken, the first company to solve this was Rosetta Stone, whose courses did not consist of memorizing vocabulary words and conjugations, but instead provided users with the context of a situation in which words were used and modeled the process of remembering language in context, comparing different situations, and assimilating language as part of the communication process.  And it was wildly successful because it solved the real problem.

And again, the real problem had nothing to do with the performance of parts of the as-is process of teaching, but with the nature of the process itself.   So long as education designers focused on improving the parts while ignoring they systemic problem, no significant progress was made.  They had to start over from a blank slate.  And because their competitors remained mired in incrementally improving a broken process, Rosetta Stone constituted an amazing leap forward.

The same is likely true of a great many products, to the extent that “new and improved” has become something of a joke.  Whenever a product bears that label, customers are dubious that the improvement is significant – they often have to search for what is new and different, and are often disappointed by what it is.  A “new and improved” detergent may have a different scent, and “Version 11.1” of a software product adds features that they have no use for anyway.  

Customers have become so jaded to this practice that in order to get their attention, a product has to launch under an entirely different brand to convince them that it is really different – and even then, there is the tendency for this fact to become known and for customers to spread the word that “new brand” is exactly the same as “old brand” (even if there are some minor differences).  Novelty cannot be faked.

Starting over from a clean slate is an exceedingly difficult process because the certainty of current practices and the fear that something different will not work out inexorably bring people back to considering incremental improvements to the as-is process.   “Let’s consider a new way to X” is followed very quickly by “We’ll start by looking at what we do today and considering ways to improve it.”  Because that’s easy, and because that’s safe.

To go a bit further, I will posit that this is the reason nothing good is ever created by a committee.  When people get together in groups, the fear of the unknown becomes a constant refrain in the “innovation” process – anyone with a bold, new vision is corralled back into the herd – and herds are characterized by a desire for safety and skittishness in the face of the unknown.  

But this is a transition to a much different line of thinking.

Organizational Disorganization

In "Anarchy in the Office" I considered the rather bizarre concept of a project execution environment in which there were no formal leaders, merely ad-hoc coordinators that leveraged voluntary resources to complete tasks.   It seemed an interesting but likely impractical arrangement, and I've since read a book that proposes to do the same not only for a specific environment, but entire companies.

Peer Leadership proposes a networked (rather than hierarchical) organization structure, in which individual employees are like nodes of a computer network that are engaged as needed to accomplish organizational tasks - or which can, on occasion take on coordination of the efforts of others when they recognize the need for something to be done.

It remains an interesting concept, though it strikes me as being even more improbable on that level, given a number of potential issues - chiefly, that most of the day-to-day operations of a business are rather routine and non-dynamic and require the ongoing involvement of the same resources with few situations in which deviation from standard operating procedure is necessary.

Granted, the problems many companies face in a competitive environment is that standard operating procedure becomes bureaucratic and inflexible, as the author rightly suggests:  a front-line employee perceives a need for a change to be made, must communicate it up the chain of command to a high enough level for someone who is not familiar with the problem to authorize a change or deviation from established process, and then the change must be socialized within the organization before communicating back down to the front lines for implementation - a procedure which can take days or months.

But as in many things, I expect swinging the pendulum to the opposite extreme could do more harm than good ... or perhaps it's just that I've become institutionalized to the traditional approach of command-and-control hierarchies that I can't fully concretize the concept.  I don't really think that's the case, but I'll allow for the possibility.

My sense is that the determination of the amount of authority given to those on the front lines of an organization largely depend on the task in question.  For some tasks (day-to-day operations that are repeated) the procedure/control structure is likely the best approach whereas for others (which focus on changing the procedures that guide routine actions) a great deal more latitude is necessary to make progress.   To impose the organizational structure of one upon the other would be counterproductive.

Measuring What Matters

At a conference I attended, there was this obnoxious little man who needled the speakers every time they mentioned a metric.  When someone cited a change in their Net Promoter Score as evidence that an initiative had been successful, he'd ask them a flurry of pointed questions about what the metric meant in terms of the firm's financial results, flustering the speaker (who was unprepared to go off on this sidetrack) and annoying the audience (who patiently waited for the speaker to be able to continue on the topic they had come to hear about).   Later in the day, this heckler took the stage to deliver his own presentation, developing the topic he had wedged into previous presenters.  What he had to say on the matter was impressive, and well worth considering.

Primarily, there is a widespread problem of metrics being chosen at a whim.  Companies latch on to a fashionable metric such as the Net Promoter Score (NPS) or Customer Experience Index (CXI) and use it as a means to measure the performance of their initiatives without, as he had aggressively suggested, considering whether there is any correlation between the metric and any behavior that is remotely of value to a firm.

That is to say that a 10% improvement in the NPS or CXI does not automatically correlate to a similar increase in revenue or repeat visits.   Sometimes, you can effect an improvement in a score while the other goes down, or achieve improvements in both scores as your revenue diminishes, your reputation is damaged, and your customers' loyalty is strained.  As such, you really shouldn't take any metric for granted.

Firms really should consider the metrics in terms of meaningful outcomes - but it's obvious that few actually do.   Management will set a random goal tied to a random metric (such as a 10% improvement in NPS) and feel satisfied when they are able to achieve the mark they set for themselves regardless of its impact on the firm's performance.   Turns out we're fond of numbers even when they don't mean anything.  Five is better than six, regardless of what is being measured or whether there's any demonstrable benefit to being a six rather than a five.

The one possible use for industry-standard metrics is to compare your firm to others in your industry.   But even that is ultimately meaningless: a company that achieves a higher score in one regard may have worse performance than another firm whose score is lower.   Doing so also means that you take for granted that the industry is doing well - to be the customer service leader in the airline or used car industry is like being the healthiest of the terminally ill.

Ultimately, the work we do is about improving the customer experience, taking it on faith that it will result in improving the financial performance of our firms - and generally, this turns out to be true.   Having a vague sense that you're doing the right thing is likely better than having no sense at all - and better still in feeling a sense of accomplishment for doing something terrible.   It cannot be taken for granted that a general metric is an indicator of success at anything but making a number better.

And in this sense, what "works" for the industry may not work for a given firm.   Two competitors may find different metrics to be meaningful:   one may find that revenues increase in correlation to NPS, another may find that revenues increase in correlation to CXI, and a third may find a completely different metric to be meaningful.

I'm left with the distinct sense that I should ask some of the same noisome questions of my own colleagues, albeit more gently and in a more appropriate forum, to determine which of the metrics is most meaningful, and encourage the pursuit only of those that can be strongly correlated to outcomes, forsaking all others for what they are: scores that are no more meaningful than a magazine quiz.

Efficiency Isn't Innovation

In general, the approach to improving things, products or processes, begins with analyzing the as-is situation and identifying areas in which problems could be fixed or improvements could be made.   That is to say that it begins rooted in present reality and ends with only minor changes.   This is different to, and likely preventative of, true innovation, which requires starting with a blank slate and imagining the possibilities that might exist, independent of what currently does exist.

It is a common, but fundamental error, to regard anything new through the lens of existing processes.  This results not in innovation, but efficiency improvements, as firms seek to streamline what they are presently doing rather than considering whether there might be an entirely new way ("new" being the "nova" in "innovation") to achieve the desired goals - or even to change the way in which the goals are defined if doing so is necessary to achieve a better outcome.

In many instances, efficiency improvements are merely automation.   In the early industrial era, automation merely replicated human motion with machines; and in the present era of information technology, automation merely replaces human thought processes with digital ones - but "merely replaces" means the that task remains the same, it is just performed by a different actor.

For example, a computerized accounting system automates the way in which invoices are processed, in that the very same thing is done with databases and spreadsheets that nineteenth-century clerks did with ledgers and quill pens.   The process is made faster, and less labor is required, but the process itself has not changed.

In that sense, replacing a worker with a machine or a clerk with a computer system is not innovative at all: it's doing the same thing more quickly and efficiently, but still doing the same thing.   To innovate requires asking: what goal are we attempting to achieve by doing things this way ... and is there a different way in which we might achieve it?"

Knowledge of existing business practices is not only unnecessary, but can be harmful.   That's not to say that they can be completely ignored - the inputs and outputs are likely still the same (though one might reconsider whether the inputs or outputs could be improved) - but all the "stuff" in the middle is entirely irrelevant.   So long as the goals of the process are achieved, the rituals by which they are pursued is irrelevant.

As a final note: innovation is not always necessary, and sometimes efficiency improvements are the best that can be done - let's not throw that concept away entirely.  But at the same time, let's not assume that the two are similar or can be accomplished in the same way.  To be innovative in the outcome requires being innovative in the process - and that holds true even when the process is one of defining processes.