Why it matters:
- Customer data will continue to drive design ideas long after the original project is completed.
- Building a full set of customer work models is an investment. You want to make it pay.
- You want to use your resources wisely. Don’t get data before you need it-but don’t get blindsided by changes in the market.
To ship a good product, make sure you understand the customers and users of that product. We’ve all bought into that message by now, I hope? And the Contextual Design work models are the most thorough and detailed method around for building and capturing that understanding.
But let’s face it — they represent a significant investment of time and effort. You’ll spend from a month to a month and a half to do the interviews, do the interpretation sessions, and build the consolidated models for a significant market. You’d like to think you won’t have to do this again on the next project, and again on the next project after that.
So how long can you keep milking the same set of models before they run dry? How long before you must return to the field to revise and extend your understanding?
When can you reuse data?
Before we go into the answer to that question, let’s look at the situations when you can reuse data. You collect data for a market from a given focus. So both the market–the people–you study and the focus you take will limit how much you can reuse data.
For a broad understanding of a market, for considering the widest and most innovative range of designs, and for the most reusability of the data, you want a broad focus. You want to focus on the work practice and how it connects across people and over time. You do not want to focus on a single aspect of the work or a single task.
The affinity and the big-picture models — low, cultural, and physical — will give you the widest range of ideas. They will nearly always suggest many more ideas and directions than you can build into the particular project you are working on. Sequence and artifact data, by their nature are more specific and targeted. Don’t expect to reuse these models to the same extent. Later product versions will nearly always be able to build on the flow, cultural, and physical models. Closely related products–products that integrate with the initial product, and products which serve the same market — can usually build on these models as well.
How do you go about reusing data?
When you are ready to start a follow-on project, use your existing work models to jump-start the process. Put up the models and affinity. Walk them asking:
- What do these models say about my market?
- What do these models suggest about my product?
- What additional data do we need not covered by these models?
It’s likely you’ll need additional sequence data to cover the tasks you’ll be supporting in the new product or version. It’s also possible you’ll need to interview additional roles for the flow model, and if you are extending your sales to new parts of the world you may want to extend the cultural model.
Tailor your data gathering to the holes you identified. Don’t worry about validating the data you have — you’ll get enough validation from the interviews you do. Instead, focus on the new roles and new tasks. If all you need to do is gather additional task data from roles you’ve already interviewed, you can do very focused interviews that just concentrate on that one task. If you gather new data, roll your new data into the existing models. Add roles to your role model; extend your cultural and physical models; and consolidate sequences for your new tasks.
This will make data gathering for later products much quicker than the first round. You’ll be able to do shorter, focused interviews with targeted users, and from those quickly develop a sound design
So, how long is the data good for?
In general, you can keep this up for five years before you have to go back and do a sweep of your market again. This number is a rule of thumb, but based on our experience it holds pretty well. Here are some examples:
We gathered data on system management from 1990 – 1995. The core of the work models we developed over this time still holds true, seven to ten years later. But we are also starting to see some differences-notably, system managers who started out with Windows servers have a different culture than the Unix system managers we studied ten years ago. They are much less likely to build special tools or tinker with the innards of the systems they manage.
To design for system managers now, we’d want to go out and study this different culture and way of working. We wouldn’t want to assume our old data holds good.
Online grocery shopping
We gathered data on grocery shopping between 1993 and 1997. From this data we developed a set of design concepts for shopping tools, home shopping environments, and the re-design of grocery stores. We did this to provide examples for our courses so we never productized any of these ideas, but the industry has slowly moved to adopt many of them on their own.
For example, we developed the concept of “safe zones” — areas within a grocery store that would be clearly distinct and give shoppers with special requirements a sense that it was “safe” to browse there. A natural-foods enthusiast would feel safe browsing in the natural-foods area of the store, for example. Just recently, the grocery store in my neighborhood installed a natural-foods “safe zone” with a dark tile floor, dropped lighting, and racks instead of aisles-right in the middle of what’s otherwise an ordinary grocery store.
Shopping itself hasn’t changed a whole lot over the past five years, but now that we’re seeing the ideas we generated from customer data showing up in the marketplace, we wouldn’t want to keep working off this data. We’d want to gather data on how these ideas are working out and take them the next step.
In the fall of 1999 we gathered data on business to business relationships. We studied how suppliers work with OEMs, how large corporations work with the banks and other entities that serve them, and how traders work with each other. We developed a full set of data characterizing B2B relationships.
Based on this data we developed a set of concepts focusing on how to facilitate the coordination and personal relationships that make B2B relationships run-even at the highest levels of the largest businesses. We saw a role for the web in making these relationships run smoothly.
Now, two and a half years later, we see the industry still struggling towards those concepts. Despite all the setbacks, it’s clear that people are still groping for ways to make B2B work better with the Web. So we wouldn’t need to start from scratch to design a web-based B2B system — we could safely go to our old data, gather some data specific to the new task, use it to validate our expectations as well and go forward.
Keep touching the market
These are extreme cases. You won’t be studying a market and then doing nothing for five years-you’ll go to field test, ship a product, ship an update, and ship the next version. All these events are excuses for going out and gathering a little more data-and this continual updating means you’ll never be really out of touch with your market. You’ll have a good sense for how well the data holds and when you need to get an update.
So don’t be afraid to commit the resources to do a quality job of understanding your market now. From the point of view of product design-and market success-it’s money in the bank.