Question 3 of 7: Radical and incremental design, and adoption (15 marks)

Question 3 — Radical and incremental design, and adoption (15 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Given. Two product introductions modelled with the Bass diffusion model: a radical change with coefficient of innovation $p = 0.02$ and coefficient of imitation $q = 0.55$, and an incremental change with $p = 0.14$ and $q = 0.30$.

Find. A reasoned classification of the 2010 iPad as radical or incremental, and a comparison of the two adoption paths — time to peak adoption rate and cumulative adoption at two years — together with the levers a manufacturer has over them.

Approach. Classify by the standard architecture-and-behaviour test rather than by technical difficulty, then quantify the adoption difference with the Bass model, whose two coefficients separate cleanly into “buys on own judgement” and “buys because others have”.

A. The iPad, when it was introduced

The useful test for radical against incremental is not how hard the engineering was. It is whether the architecture, the working principle and the user's behaviour all change together. By that test the 2010 iPad was radical at the level of the product category and the business system, while being almost entirely incremental at the level of its components — and holding both halves of that statement at once is the answer the question is asking for.

Almost nothing inside it was new. Capacitive multi-touch, ARM systems-on-chip, lithium-polymer cells, backlit LCDs, accelerometers and Wi-Fi radios were all mature, and Apple itself had shipped the whole stack three years earlier in the iPhone. Tablet computers had existed for a decade in the Microsoft Tablet PC form. On a component-technology reading the iPad is an incremental scaling of an existing product.

What was radical was the system architecture and the definition of the category. Earlier tablets were personal computers with a stylus bolted on: they ran a desktop operating system, a desktop file system and desktop applications, and they were therefore worse laptops. The iPad discarded the desktop paradigm entirely — direct finger manipulation with no pointer and no stylus, one full-screen application at a time, no user-visible file system, instant resume with no boot, all-day battery life bought by a mobile SoC, and software distributed only through a curated store with the payment relationship already established. That last element is the one engineers most often miss: the App Store was as much a part of the design as the enclosure, because it turned a device with few uses at launch into a platform whose uses grew without Apple's involvement. Consumer behaviour changed accordingly — media consumption moved to the sofa, and a device with no keyboard became the primary computer for a large population that had never wanted a computer.

The honest classification is therefore: radical innovation, incremental technology. It is an architectural innovation in the Henderson–Clark sense — known components recombined into a new architecture and a new use pattern — and it is a useful corrective to the common error of equating radicalness with technical novelty. A twelfth sprocket on a bicycle cassette is far harder engineering than bolting a motor to the frame, but the electric bicycle is the radical change, because it alters what the object is for and who will use it.

B. How adoption differs, and how it can be influenced

Adoption differs in shape, not merely in speed, and the Bass model makes the difference explicit. Cumulative adoption is

$$F(t) = \frac{1 - e^{-(p+q)t}}{1 + \dfrac{q}{p}e^{-(p+q)t}}$$

where $p$ is the coefficient of innovation — the propensity to buy on one's own assessment, independent of how many others own the product — and $q$ is the coefficient of imitation, the propensity to buy because ownership is visible in one's social network. A radical change is characteristically low $p$ (few people can evaluate a category that does not yet exist) and high $q$ (once a friend owns one and can demonstrate it, the argument is settled in a minute). An incremental change is the reverse: buyers already understand the category and can judge the improvement themselves, so $p$ is high, while there is little social signalling in owning the newer version, so $q$ is low.

The inflection point of the curve — the moment when the adoption rate peaks and, in practice, when capacity must already be in place — is

$$t^{*} = \frac{\ln(q/p)}{p+q}$$

For the radical case, $t^{*} = \ln(0.55/0.02)/0.57 = \ln(27.5)/0.57 = 3.314/0.57 = \boxed{5.81\ \text{years}}$, whereas for the incremental case $t^{*} = \ln(0.30/0.14)/0.44 = 0.762/0.44 = \boxed{1.73\ \text{years}}$. Evaluating the cumulative fraction at two years gives

$$F_{\text{rad}}(2) = 6.9\ \%, \qquad F_{\text{inc}}(2) = 31.0\ \%$$

while by six years the ordering has begun to close, at 50.9 % against 80.5 %. This is the quantitative statement of a familiar commercial trap: a radical product looks like a failure for two to four years on the same evidence that an incremental product uses to declare success, and the temptation to cancel it is strongest immediately before the curve turns. It is also why the peak adoption rate, $(p+q)^{2}/4q$, is 14.8 % of the market per year for the radical case against 16.1 % for the incremental one — similar peaks, arriving four years apart, which is a capacity-planning problem as much as a marketing one.

How adoption can be influenced. The two coefficients respond to different actions, and the design decisions that move them are made long before launch.

To raise $p$, the manufacturer must reduce what the first buyer has to take on trust: reduce the perceived risk with trial, generous return policies, warranties and in-store demonstration; reduce the switching cost by designing for compatibility with what the buyer already owns (Rogers' compatibility attribute); increase observability and trialability so the benefit can be seen in thirty seconds rather than explained in an hour; and reduce complexity, which is the strongest single design lever — the iPad's zero-instruction interface is precisely an attack on $p$. Price is a lever but a weak one for radical products, since the barrier is comprehension rather than affordability.

To raise $q$, the manufacturer must make ownership visible and demonstrable: products used in public diffuse faster than products used in private; a distinctive form makes ownership legible at a distance; seeding to opinion leaders and lead users creates the first demonstrations; and network or platform effects — the App Store again — make each additional owner genuinely increase the value to the next. Referral incentives act directly on $q$.

Two cautions. First, actions that raise $q$ before the product is good enough are actively harmful, because word of mouth carries a bad experience just as efficiently as a good one. Second, the strategic choice implied by the model is about capacity and cash: a radical launch should be planned with modest early capacity, a financing runway that survives four flat years, and a supply chain able to scale sharply at $t^{*}$; an incremental launch must have full capacity at day one, because its peak arrives at 1.7 years and a stock-out at the inflection is unrecoverable.