Do Phase—Experiment like a Lab Technician
The proof of the pudding is in the eating.
—William Camden
Making your ideas concrete is what this section is all about. You will learn how, through smart design, you can offer your internal or external customer precisely what they need, and how to actually build this design. You discover the usefulness of a minimum viable product and how you can test it properly.
12.1 Building Customer-value Propositions
You’ve put enough thought and planning into the Think phase. Now it’s time to go to work and prove that your ideas actually work in practice. In other words, prove that your internal or external customer will purchase or use the product or service you create, or value the improvement(s) you make. So you must offer that customer something he appreciates. Whether that is an improved assessment-process in HRM or a new search option in your app, the principle remains the same: how do you offer added value? This concept is called a customer-value proposition. Once again, the design and construction of your proposition will benefit from a visual approach, using brown paper, Post-its, and markers. Figure 12.1 shows a format (which comes from the excellent book Value Proposition Design by Osterwalder,1) which you can put on a whiteboard or brown paper and the one shown is for a vendor-neutral fuel card.
Figure 12.1 The customer value proposition format
The methodology revolves around the persona (see section 10.3). In the example, these are freelancers and other small-business owners. Next, step-by-step, you answer the following questions:
- What tasks must the customer perform within the framework of what you can and want to offer him? (See also Section 10.3.)
This entrepreneur has to fill his vehicle(s) with fuel and keep business accounts for this process.
- What emotional and rational benefits would the customer thereby gain?
He wants to deduct sales taxes and business expenses. Because he is busy, hates bookkeeping, and often makes administrative errors, it must be quick and easy. He also wants to minimize the costs, be able to refuel anywhere and take advantage of loyalty programs.
- What problems, inconveniences, risks, and barriers might the customer experience?
He would fail the credit-check process of the big fuel brands if he applied for their fuel cards. Therefore, he uses his debit card and processes the receipts in his administration. But he often loses receipts and that is not accepted by the tax authorities, so he runs the risk of fines. If he would have a brand-dedicated fuel card, filling-up at another petrol station would just bring similar hassles.
- What characteristics have your products, services, and channels?
A universal fuel card that is coupled to online administration and a supporting app.
- What is the alternative?
A fuel card for the major fuel brands. Or use a private car and claim business mileage.
- What advantages do your features provide the customer?
The pass can be used at all petrol stations. The customer can easily see which are the cheapest garages. Application and approval are possible 24 hours a day online. Direct debit ensures he always pays immediately. The pass can also be used for parking and the car wash.
- How do the features solve or ease his pain?
Online, automated financial administration ensures his sales tax and costs are always deductible and prevent personal errors. He need not perform time-consuming operations. Furthermore, he is not tied to the sites or the high prices of major fuel brands.
As is apparent from the example, there should be a “fit.” The better you can meet the need, the more successful you will be. This is at the core of everything, and you can apply this at any level—to considering a new type of aircraft such as the Dreamliner, developing the first Tesla model, inventing the Netflix service, or expanding payment functionality in a shop. And it gives you a base to optimize continuously. With the fuel card in the example above, you could add functionality with an app that automates the keeping of mileage records, so avoiding additional tax liability. Or a solution for automated-payment when using public transport even only rarely.
A similar development-process is used by Ikea. For over sixty years, Ikea has been a distinctive pioneer in the home furnishings market. Firstly, because the company broke up the traditional market segmentation: buyers of expensive furniture or buyers of cheap furniture. Ikea saw that design need not be expensive, that the combination of cheap and expensive furniture would become a trend and that your life situation affects your choices: whether you are about to study, live together, get children, or buy and furnish a second home. Secondly, Ikea introduced self-service, meaning their furniture was not only cheaper than all their competitors, but could be taken home immediately. Scientific research shows that self-assembling furniture ensures customers better appreciate their purchases.2
But, around the turn of the century, Ikea noticed a change. Some customers had no time to assemble their purchases or were not proficient enough and, therefore, found it convenient to have someone assemble it for them. So the company piloted an assembly service which, after some tweaking, was rolled-out across all locations. In addition, Ikea noticed some customers could not transport large items themselves. So Ikea began renting trailers and also offered a delivery service. You’d think they would stop there, but they were constantly looking to improve the customer experience and so increase sales. By walking around the stores, management saw there were also customers who hate shopping, had not enough time for it or simply were not good at finding what they were looking for in the enormous warehouse areas. So Ikea added a service: customers could submit a wish list, which Ikea staff would collect and arrange delivery. And just recently (because Ikea knows how many impulse purchases occur in physical stores) Ikea combined these services into its e-commerce. You can now buy 9,000 Ikea products online, and have them delivered to your door, and have them assembled if you wish. Moreover, Ikea had already, in the late 90s, made headway into the business market with similar solutions.
Back to your customer-value proposition. Now the foundations are in place, you need to make sure these manifest in the customer experience. Read more about this in the next section.
12.2 Building Customer Experiences
As the SERVQUAL model makes clear, achieving customer satisfaction is, above all, a matter of matching the expectations of the customer with their perceived experience. In addition to having the right core products and services, it appears that the customer experience—at all the organization’s touchpoints—also plays a vital role3: the customer’s actual experience of the proposition, as described in Section 12.1. And that is something your organization can build-on for continuous improvement. But what is a customer experience exactly? Let’s look at an example from retail.
Imagine you are in a shoe shop. After a long search, you’ve finally found the perfect shoes, but the last pair in your size has just been sold. After a deep sigh, you walk over to the cashier and ask if there’s a pair in another branch. Like a kind of Sherlock Holmes, the assistant calls one branch after another, searching for the right size. Eventually, you leave the shop, empty-handed, with the vague promise that they will call you. Frustrated, and doubtful if you’ll hear from them, you start looking for another shop to begin your search anew. And there goes another chunk of your free time.
At Decathlon, this experience is definitely a thing of the past. In their 1,120 stores worldwide, there are large touch screens with access to their webshop. Here it takes a moment or two to sign up and then you can immediately see the actual stock of all their stores and their central warehouse. You order the product, the machine prints the invoice, and you pay at the checkout. Delivery is free and you get a confirmation mail with your Track & Trace code. Convenient, right? This is useful even when the product is in stock in the store, but you want to continue shopping without having to lug a heavy bag around. Or if you want to see the much larger range in the webshop. Or if you want to receive advice from one of the assistants about products that are never in the shop, such as large fitness equipment. And you can instantly make and purchase personalized products, like a football jersey with your child’s name on the back.
With several companies, such as Three (one of the UK’s leading mobile operators), it works the other way round: they use an offline channel to improve the online experience. On the site, they meticulously follow your behavior and use these data to predict what you are likely to need. If, for example, you are finding it hard to choose between two subscription packages and clicking from one page to the other, you will get a pop-up screen asking if they can help you directly; for example, by live-chat or phone, after you enter your mobile number. An employee can then guide you to the best choice, by asking you questions and giving expert advice. Customers seem to appreciate this enormously.
Sounds great right? Perhaps you are wondering how Decathlon and Three have so nicely thought through all the issues in detail, and how, perhaps, you could do something similar for your own customers’ experiences. For the answer, let’s take a trip to the world of film: you are going to use a method called storyboarding to describe scenes from the proposed experience. And while you might feel like a third-rate director of a B-movie, you can still put together an awesome script. Everything the customer should experience must be there, obvious, simple, and without nuances. The elements should be clear to everyone in your organization, so they know exactly what is expected of them. Now you can begin thinking through the elements, with your team—a brainstorming session—with brown paper, white boards, Post-its, and marker pens.
Figure 12.2 shows a simplified storyboard for an internal customer: a participant in a training course organized by HR. As you can see, the experience is cut up into individual scenes that are described vividly, using colorful imagery. YouTube has a host of funny negative examples we can all relate to; search for “analytics in real life” or “soup nazi.”
Figure 12.2 The customer experience of a participant in an in-company training
The elements that emerge in your scenes are the blueprint for creating a concrete customer experience. But you still have to answer the last question, from the business model canvas in Chapter 10, about revenue streams:
- Which features and benefits of our value proposition and channels are our internal/external customers really willing to pay for?
- What are they paying for at the moment? And how do they pay?
- How would they prefer to pay? Consider the following: price per unit of product or service; user fee or licence; subscription; rent/lease; hourly rates; mediation costs.
- Which pricing mechanism is the best fit, according to the customers? Consider: suggested/list prices; price per quality level or optional product characteristic; volume levels; negotiated price; auctions; real-time market (like shares); yield management (such as airline seats).
- How does each revenue stream add to the total revenue?
Now that you’ve finished your design, it’s time to start building. And your tool is the smart approach we met earlier, the minimum viable product. How to use it is the theme of the next section.
12.3 Learning Faster and Cheaper with the Minimum Viable Product
If you buy a new car, do you buy it immediately or do you take a test drive first? And if it is a used car, do you read the service book or maybe even have the car checked, or do you just assume that it’s all right? When you are buying an existing home, do you first visit it and perhaps commission a structural inspection, or do you just buy it directly via a “property portal” such as Zillo and Zoopla? Do you book a hotel or holiday home directly from the site of the travel agent or do you first look for local information and reviews? And what about that interesting new drink that you haven’t tried yet? Do you knock it back in one gulp up or start with a careful sip? And the shower or bath? Do you step right in or first put in a hand to feel how hot the water is? Would you immediately marry that (potentially) great partner you met, or do a bit of dating and live together for a while before making such a big decision?
First Try, Then Do
Most people will first perform a kind of test to determine if something comes close to meeting their expectations (SERVQUAL again). It is natural behavior that has been deeply rooted in us for millions of years. We try to eliminate risk by trying something out carefully and if it proves to be safe, we continue. Just as animals do: first sniff a bit and then take a bite. An instinct that helps you survive.
The strange thing is that in organizations, it often works differently. There are countless examples of investments where a lot of money was put in, only to find, long after at the launch, that the product is not working properly or that customers do not need it. As you read earlier, this way of proceeding is called the “waterfall” approach. This occurs internally, such as in IT projects in government institutions, but also externally. The latter is most evident in products and services for consumers. Take for example the market introductions New Coke, the BMW C1 scooter, Persil Power, and computer-voting. There is even a museum of failed products. Annually, an estimated 156,000 new products come to market, a new-product launch on average every three minutes. And 76 percent fail.4
This return may be higher, but for that we need to transcend alchemy. That organizations innovate is, of course, very nice. But when intelligence and structure are brought into the experiment, the results can be improved—with a predictability which would make a match fixer very happy.
Reviewing Your Production
Agile management takes a logical approach to improving results. In Chapter 11, you read all about the story and the definition of done. These serve to deliver, at the end of a sprint, an improvement that works and has been tested “technically.” To determine whether it does do what it is supposed to do, the team performs an inspection with the product owner and any other interested parties. This informal meeting is called a review and lasts not more than two hours. The aim is to promote cooperation by looking together at what has been accomplished during the sprint. On the basis of their feedback and the ideas generated during this meeting, the value of what has just been made can be further increased, and the product-backlog immediately adjusted.
Another important success factor in the approach is the minimum viable product (MVP). Let’s look at what that means.
MVP: Back to Basics
The MVP (serial-entrepreneur Steve Blank calls it the minimum feature set) is a concept which allows you, with the minimum possible cost and at the highest possible speed, to learn how much your internal or external customers appreciate your products, services, and experiences or the improvements you’ve made to these. The simpler the better (like the story in section 11.2). Instead of spending a lot of time and money trying to create a perfect, ready-to-use product, you strip away everything, except what is absolutely necessary, until you’re left with something that is just about viable. So you can get faster feedback from your customers and avoid unnecessarily investing in tasks to which the customer attaches no value (the Lean value-creation which you read about in paragraph 7.3). So you first try-out what you’ve made before you improve it further, so as to avoid wastage. In other words, you can smart-test the hypotheses that you generated in the Think phase. That is the purpose of the MVP and fits perfectly with the rapid iterations of the Think–Do–Learn cycle.
The limits to the viability of an MVP are determined by the customer: which quality level is acceptable to him, in terms of completeness and predictability of performance? Conversely, you want to learn as quickly as possible, plus you want to work as cheaply as possible in terms of the amount of time and money you invest. These three factors give you a dilemma, as you can never score well on all three criteria. You have to be smart here and look for the best balance (see Figure 12.3).
Figure 12.3 The devilish dilemma of the minimum viable product
How does this work exactly? Architects, for example, work with 3D software and models to talk through their designs with a client. Film studios show animated storyboards, and sometimes rough trailers to panels of frequent moviegoers. Car makers create mock-ups of studio models and show these concept cars at trade shows to gather reactions from media and potential customers. Comedians first do tryouts in small venues before they go into a real theater. Fashion designers create sketches of their clothes and match them with fabric samples to demonstrate their new designs to the decision makers at the big brands.
The idea is that, via the MVP, your customers get a good understanding of what you want to offer them eventually, an unusual application of fake it until you make it. This seems to work best with early adopters, a term coined by Everett Rogers.5 Early adopters are lovers of innovation, they like to try new things, and so are open to an MVP. They are visionary enough to imagine the missing features themselves, if necessary. Another advantage is that you can also use early adopters to engender enough curiosity in the early majority to make them want to try your proposition out. This effect is called crossing the chasm6: bridging the critical gap to the practically minded and more-expectant masses.
What shape can your MVP take? Here is a nonexhaustive list of commonly used forms suitable for internal and external customers:
- Idea or concept—If you have an idea or concept in your head or have put one on paper, for example using the right-hand section of the business model canvas, you can investigate this with your (potential) customers in key interviews and focus groups (you could go even further by running co-creation sessions.) Or you can discuss the idea or concept via your own blog page, so you can gauge the reactions of visitors and get into discussions with them.
- Mock-up—Here you take the next step and let your (potential) customers see your concept: it might be a clay model, a mood board, or a wireframe that you show them. Or a landing page on which you show fake product images or screenshots. When visitors click through for more information or to order, they get an error message, or a message that it is temporarily sold-out or otherwise unavailable. Or they come to a page where they can sign up for a (free) trial, or can subscribe to get an email as soon as it is available, or even order the product in advance. You attract visitors to the landing page by promoting it on your website, or via a Google adwords campaign, or a (viral) campaign using social media.
- Demo—Going one step further, you can offer a demonstration of your service/product, by putting an explanatory video on the landing page (this is how Dropbox started), or create a fake site where you click through and make it appear to really work.
- Prototype—The next step is a working beta-version of the proposition, a one-off makeshift product a bit like a patchwork quilt: a bit from here; a bit from there. To do this, there are all kinds of cloud solutions available: platforms like Amazon Web Services and Google Forms. Sometimes, you have to do some manual work behind the scenes, which will ultimately be automated. You can choose to imply that it is all automated or you can be fully transparent to your users. Zappos.com, for example, started with the first approach. Finally, you can also choose not to display “all” your features in limited form, but only one more or fully developed feature, the most important of course. And here Google is the killer example.
The MVPs discussed above show that it often involves a combination of the form of the product and the test method. However, some test methods are suitable for multiple forms and some forms are suitable for multiple testing methods. Therefore, it is good to look at the test methods separately. This we do in Section 12.4.
12.4 Testing: to Measure Is to Know
Imagine that you are working for a sports brand, which has been producing sports footwear for more than forty years. You notice a retro trend in the market and wonder if there might be demand for a vintage model. But you know that it would be far too expensive to just take the gamble and produce a series. That old chicken-and-egg problem. How can you solve this? By going to your product archives, selecting some old models, photographing them, and putting them on eBay. If enough people look at them and place bids for them, you will get a good indication of the potential. If it is large enough, you decide which models to get back into production. That is the way Nike did it.
Another useful insight can be when you realize that there are people who believe enough in your product or service that they are willing to invest in it. And these are not professional investors, but people who could also be your clients; crowdfunding. There are numerous platforms where you can organize crowdfunding easily and quickly, such as Kickstarter; if it works, it also increases your working capital.
What the above makes clear is that within agile management, there are actually two types of tests. As Section 11.3 showed already, you first do a technical test on what you have built. The aim is to determine whether what you’ve built works. In addition, as the Nike and crowdfunding examples above demonstrate, you also want to test its customer potential. This type of testing we discuss below.
Commitments rather than Words
So what you do, in this type of test, is look at how customers and prospects respond to what you have designed for them, whether it is a tangible product, an online service, or an improvement in the functioning of a specific channel. And as much as you might do some exploring in the initial phase, by asking customers about their needs and preferences, you really want to know if they will, eventually, actually go out and buy or use what you are offering them.
You immediately come up against a problem which, in 1962, David Ogilvy identified: “People do not think what they feel, do not say what they think and do not do what they say.” Since then, this has been confirmed by neurological research. Scientists might argue about the exact percentage, but 90 to 95 percent of our behavior seems to be unconscious. Functional MRI (fMRI) scans provide hard evidence of this phenomenon: our brain has made a decision for us on average seven seconds before “we” think we make that decision.6 These unconscious decisions, in practice, appear very reliable, so researchers use them as a basis for predictions. When an fMRI of people seeing, for example, three different Cosmopolitan covers, shows that one version generates the most buying impulses, it also later appears, by far to do best in newsstand sales. Yet the research subjects themselves cannot say exactly why they chose that particular cover.
For example, you want to buy a house and you are going to look at some houses. As soon as you put your foot on the garden path, your brain has already decided if it is going to be this house or not. But you still sit down afterwards and try to make a rational decision, for example by drawing up a list of pros and cons, which you might even weigh and score. Very objective, right? Not so. Your brain continues, unconsciously, to control the selection, weighting and scoring your arguments so that your results seamlessly match what it has already decided for you. Your brain makes everything fit in with what has already been established, in order to prevent cognitive dissonance. However, we think of ourselves as rational beings and classic market research is based on this idea. You talk to customers, using surveys, focus groups, and interviews, mostly asking about intentions for future behavior in hypothetical situations. But the effect of unconscious behavior calls into question the predictive value of this approach.
What is the best way to test if the results do have a high predictive value? Not with words but with deeds: you shouldn’t talk to your customers, but observe their actual behavior, and not like an old-fashioned spy behind a newspaper with two eyeholes cut into it. Fortunately, there are much better sources of information, as you will see in Chapter 13. But in the next paragraph, you can already find out more about the most important source.
The AB(C) Test
For many organizations, it is a challenge to test in a pure and objective way. The first and major obstacle is what they call, at Google, the “HiPPO” effect (Highest Paid Person’s Opinion): teams allow their decision-making to be influenced, too much, by the opinions of their leaders. And therein are great risks. It is vital that your conclusions are as factual as possible (see Figure 12.4). On the podium of analysis, facts get the gold medal, while arguments, experiences, and guesstimates get silver. And in third place are assertions and opinions, which you would be better off disqualifying and taking the bronze away. As Deming said: “In God we trust, all others bring data.”
Figure 12.4 Focus in testing as much as possible on hard factual data
So think of yourself as a chemist in a laboratory, and work as objectively and as structured as possible. Precision is essential. Your hypotheses from the Think phase, which you obviously did not formulate for nothing, should now serve as the basis for what you are going to test. And an important principle comes into play: ceteris paribus—and that’s not one of Harry Potter’s magical spells. The literal meaning of this Latin phrase is: with other conditions remaining the same. In other words, we try to change just one thing at a time, so that we can measure its impact, isolated from and excluding everything else. Difficult? It turns out to be easier than you think.
This is really well suited to an A/B-test, aka a split-run test (see Figure 12.5). Suppose you want to change the packaging of a product to increase sales volume. So, of course, you want to find out what adjustments work or not. In half of the selection of representative stores, you place the new packaging, of which one aspect has been changed, for example, the background color. In the other half of the stores, you leave the existing packaging. If the new packaging appears to attract higher sales than the old, you know the background color is a success factor. And that is possible because you can make a comparison with what is called the “control group.” This approach can be applied anywhere. To the format of a call-center script, a font in your application, the place of a button in your webshop, the order of an online form, traffic routing in your store, and so on. You can also expand the number of variants, for example to three, and that’s your ABC test.
Figure 12.5 The principle of the A B testing
Now that you’ve built and tested your adjustments, it is time to evaluate the results. You want to know whether the implementation of your plans and the ensuing results match your hypotheses. You can read more about this in Chapter 13, which deals with the Learn phase.
By reading this chapter, you’ll have discovered the following:
• You can make your ideas for changes and improvements concrete, by working them into a customer-value proposition. In it you translate the benefits sought by your internal and external customers, and the problems they experience in performing their customer jobs, into a suitable solution.
• In order to design and build the ideal customer experience of this solution, you can employ the storyboard approach.
• To find out quickly and cheaply if your solution is relevant and functioning properly, you can use a minimum viable product (MVP). This can take the form of an idea, concept, mock-up, demo, or prototype. The MVP is actually a form and test method combined.
• The testing is done on the basis of the hypotheses from the Think phase. With your choice of test methodology, it is important above all to observe the actual behavior of your internal or external customers, and to ensure you perform your measurements in a ceteris paribus situation. An A/B test fits perfectly here.
References
1. Osterwalder, A., and Y. Pigneur. (2014). Value Proposition Design. Hoboken: Wiley.
2. Norton, D., and D. M. Ariely. (2012). The IKEA Effect: When Labor Leads to Love. Journal of Consumer Psychology 22–23, pp. 453–460.
3. Service 2020: Return on service. (2014). Economist Intelligence Group.
4. Lindstrom, M. (2009). Buyology. New York: Random House Business.
5. Rogers, E. (2013). Diffusion of Innovations. New York: Free Press.
6. Moore, G., and R. McKenna. (2009). Crossing the Chasm. New York: HarperCollins.