Practice

Archived Posts from this Category

January 18, 2009

Model visioning

Posted by bridgeland under Model visioning, Practice
1 Comment 

In the book, we describe how business models are useful for persuasion, for changing someone’s mind about something. Of course business models have many other uses as well: analysis, training, etc. Seven other uses are described in the book.

Among business models, business simulations are particularly persuasive. In the book we say:

… simulation can be a powerful tool for persuasion, and this use of simulation is not widely appreciated. In fact, simulations are used so rarely for persuasion that sims are something of a secret sales weapon, an advantage to the people and companies who make use of them.

BJ Fogg, the director of the Stanford Persuasive Technology Lab, says, “Cause-and-effect simulations can be powerful persuaders. The power comes from the ability to explore cause-and-effect relationships without having to wait a long time to see the results and the ability to convey the effect in vivid and credible ways.” [Fogg 2003] Today when people want to persuade, they typically uses verbal techniques, numbers and graphs, or images and video. All these media have limited effectiveness. People dismiss words, ignore numbers, and are sophisticated critics of images and video. But simulations give them the ability to try things out, to experiment and build up their own understanding inside the simulated world. Simulations encourage people to reach their own conclusions through their own trial and error, but faster and more safely than they can in the real world.

[From the book Business Modeling: A Practical Guide to Realizing Business Value, by David M. Bridgeland and Ron Zahavi, published by Morgan Kaufmann Publishers, Copyright 2009 Elsevier Inc. All rights reserved.]

Let’s look at an example. Suppose you are attempting to sell Indian offshore call center services. Like most offshore call centers, yours provides telephone-based customer service cheaper than can be done by onshore call centers in North America. The main value proposition is price. Of course price can be easily understood without a need for simulation.

But the situation is a bit more complex. You differentiate yourself from other offshore call center services (who also offer low prices) by your relentless focus on accent. Most offshore call centers speak Indian English, but your call center employees speak American English with a South Carolina accent, thanks to hundreds of hours of training, monitoring, and coaching. You even focus your hiring on dialect. Your call center is based in Mumbai, home of Bollywood, the Indian movie industry. You hire aspiring and out-of-work actors and actresses to fill your cubicles, hiring employees who delight in playing characters from Charlestown and Greenville.

But so what? Why should your clients care about accent? There are several business advantages that accrue to your clients. First, some residential callers from North America have difficulty understanding Indian English. Communication with someone speaking American English is easier and more natural. Second, some residential callers from North America become uncomfortable when they realize they are speaking with someone from the other side of the world. Many people have not realized that telecommunication is free, and have little day-to-day experience with international calls. These callers never realize they are speaking to Amita and not Amy. To someone from Ohio or Oregon, South Carolina feels familiar to them; Mumbai does not. Finally some residential callers are politically opposed to speaking with Indian call centers. They see call centers in India as the “export of American jobs”, and they become angry when they talk to someone speaking Indian English. When they talk to your staff, they are usually fooled into thinking the call center is really based in South Carolina.

You have a compelling value proposition for your call center, but how do you show that value prop to your prospective clients? Of course you provide demonstrations. You play recordings of calls between North Americans and your call center, to show how natural the conversation is.

In addition, you provide a simulation, a web-based strategy sim that allows your prospective clients to try different alternative strategies. Working in the sim, your clients can experiment with different call pickup times, how quickly the calls are answered. They can experiment with different targets for time spent on the level 1 calls. They can experiment with different strategies for when to escalate to level 2 and to level 3. And of course they can experiment with different degrees of accent and dialect.

Their experiments have results. Your clients can see the results on time spent, on costs, on success rate for solving their customer’s problems, and most importantly on customer satisfaction. They can see for themselves the big and lasting effect accent and dialect has on customer satisfaction. And they can experience for themselves the difficulties of making an offshore call center work when customers are uncomfortable or angry. (Customer satisfaction is a “soft variable”. Soft variables will be explained in a future posting.)

You have built a model visioning sim. Model visioning is the use of business simulation for sales pursuits.

Model visioning is most useful for big sales, sales of a $100 million system integration project, or a $200 million outsourcing contract, or a $300 million construction plan. These big projects have big business consequences, positive and negative. Anticipating and understanding all the consequences is hard. Buyers need help in thinking through the big project. They need a custom-built business simulation to experiment with.

Traditionally what do people do to sell big projects? Big projects are sold with meetings and presentations, long written proposals, and demonstrations. A model visioning sim does not replace any of these techniques. Instead it complements these techniques. At the meeting, the sim is presented. Since it is a websim, the URL is provided to the prospective client. He can try it on his own, experimenting with different approaches to the project.

Model visioning is new: few people have employed sims in the sales process. It is virgin territory. When I create a sim for a big pursuit, and show it a client, I feel like the Dutch explorer Aerjan Block seeing Long Island for the first time. Who knows how big this is, or how important?

 

December 10, 2008

Web-based sims, the late great BellSouth, and Forio

Posted by bridgeland under Practice, Tools
1 Comment 

In the book, we describe business motivation models, and how those models can be simulated. When you simulate a motivation model, you can try different strategies, and understand the potential outcomes of each. A “strategy sim”—the simulation of a business motivation model—is a particularly effective tool for training, for bringing a group of people to a new understanding of the dynamics of a business.

But how should a strategy sim be delivered to the people who want to use it? Strategy sims want to be delivered as web applications, with a user pointing his browser at the right sim URL, trying a variety of different potential tactics, and exploring the resulting impacts on his business. The web makes business sims easy to access, no harder than looking up something in Wikipedia or buying something on Amazon.

I became convinced of the value of web-based strategy sims in the mid-1990s, while performing some management consulting work for BellSouth in Atlanta. As you may know, BellSouth was one of the “baby bells” resulting from the breakup of AT&T in 1984. BellSouth was based in Atlanta, and offered local telephony to customers in Florida, Georgia, North Carolina, and other states in the southeast United States. BellSouth has since been acquired by (the new) AT&T.

In 1995, my client at BellSouth was interested in conducting a two day training seminar for all managers in the company. The purpose of the seminar was to introduce the BellSouth management to some of the changes taking place in the telecom market, and to prepare the company for some coming transformations in their business. The seminar was centered around a playable simulation, a strategy sim that reflected the ideas of the BellSouth thought leaders.

We created this sim for them. In the sim the user could make investments in consumer broadband (then a radical idea), make pricing decisions for long distance (then a rather expensive service, not bundled for free with landlines and wireless as it is today), and other telecom decisions. The user simulated forward for six years, from 1996-2002 (!). Each year the user could investigate what had happened, make decisions, and then advance to the next year.

The sim was used competitively. The participants were divided into teams. Each team explored what had happened, and discussed what should be done. The teams competed against each other, each trying to create the best end result for the simulated BellSouth. And the competition engaged them: they wanted to beat their friends and colleagues, and many spent extra hours in the evening learning the sim, and in the process, learning about changes in the telecom competitive landscape.

In creating the sim, we focused on the user interface, making it very attractive and intuitive. Actually my old friend Nadine Carroll did all the heavy lifting on the user interface. She wrote it in Visual Basic—an obvious choice in those days. Ed Vail and I wrote the model using the strategy simulation tool Powersim. To the user it was a standalone Windows 3.1 application, with the model and simulation engine completely hidden inside the app.

For the training seminar we needed the sim to run on 150 laptops. BellSouth supplied the laptops, shipping them to my office in Boston. Nadine had created a installer application with all the necessary DLLs. Of course this was before CDROMs were common, so the installer was delivered on a series of five diskettes. We copied the diskettes, then set up an assembly line to install the app on all the laptops.

What a pain! And it did not even work right. The laptops were not identical, and the app worked on most, but failed on some. The root cause was a DLL conflict (of course), and Nadine had to figure out the offending DLL and devise a workaround.

While installing the sim on the 150 laptops, I began to fantasize about a different architecture. What if the sim were a web application instead of a Windows application? The user interface could run in a web browser, and the simulation would run on a server. No installation would be required. The user would point his browser at the right URL and play the simulation. The server could even manage the competition, tracking the high score and which team had achieved it.

The next year I joined Powersim Corporation, and we created the product Powersim Metro to achieve that vision. Metro allowed someone to create a simulation and deploy it as a web application. Metro was not a great success as a product; it was a typical version 1.0 product that partly achieved its objectives and partly missed. And it sold poorly, even for the smallish niche of modeling tools.

My Powersim colleagues Michael Bean and Will Glass later created a company that has enjoyed some success with this kind of web-based simulation environment. Their product Forio Broadcast allows you to create a simulation with a web user interface. The model is simulated on a server, either on one of Forio’s servers or your own. The user interface is written in some combination of HTML, Javascript, and Flash, and runs in a browser. The result can be a sim that feels very smooth and natural, easy to access from anywhere.

 

November 28, 2008

Model credibility and transparency

Posted by bridgeland under Practice | Tags: |
Leave a Comment 

Business models often elicit skepticism. “What is behind the model?” someone asks. Behind their polite question you can hear some less polite variants, questions they are thinking but do not say: “Why should I trust the results?” “How do I know you haven’t cooked this model for your own benefit?” Even: “I will not be fooled by your scam.”

Model credibility is often a problem. We are successful with our business models only to the extent that other people trust those models. And frankly their concerns are often warranted. As modelers, we know that all models are wrong. Some models are wrong in ways that undermine their conclusions. And some models are indeed cooked—not ours of course, but those created by less scrupulous modelers.

Model credibility is a fundamental problem with all kinds of modeling, not just business models. When we create a model we understand the details. The consumers of models do not understand the details. There is a fundamental tension between our deep knowledge and their lack of it, a tension that plays out in their skepticism.

This is similar to what happens when I drive a car for three years and then sell it to someone else. I know the defects. I know whether I crashed it on an icy road last January. I know whether I forgot to change the oil, and how aggressively I drive when listening to Big D and the Kids Table. My would-be buyer knows none of this. He reacts with skepticism.

To economists the used car sale is an example of information asymmetry. The seller has more information about the product than the buyer, and better information as well. The information is asymmetric.

Business models suffer from information asymmetry just as used cars do. Those of us who sell business models know more about the models we build. Our buyers know less. The credibility problem is inherent in this asymmetry.

What can we do to gain credibility? Some people suggest model validation (or verification). (We describe the validation and verification of business models in our book.) I think these people are mistaken: validation does not help. Validation and verification are techniques to make a model more accurate, a better reflection of reality. But the credibility problem is not about the model, it is about how the model is perceived. A skeptic can be just as skeptical about a more accurate model as about a less accurate one.

So what can we do? In practice, we use two approaches. First we engage the model consumers in the modeling process. We work with them in model-based workshops to create the model. In essence the model consumers become modelers, albeit modelers without the technical modeling skills. People who help us create models are never skeptical of the results.

Of course there are limits to engaging the model consumers. The workshops for creating the model rarely include more than nine or ten people. Everyone else is outside the room. And often when the model is created, we do not even know the ultimate consumers. Their experience with our model is later, only after the model is finished.

The second approach to creating credibility is to make the model transparent. Everything is shown. Everything is made clear and easy to understand. A model consumer who is sufficiently motivated should be able to open up the model and understand everything.

Making a transparent model is hard, harder than making an opaque one. Models often include messy details. These details must be cleaned up so they can be shown. Models often have complexities. These complexities must either be simplified, or at least explained in a simple way.

Making a transparent model is indeed hard, but it is worth the trouble. When someone is skeptical we show him the model and invite him to investigate.