IBM’s supercomputer was developed to give questions for answers that humans can understand in natural language.   They can be very difficult for a machine, however.   Answers such as “unearned Nobel Peace Prize” – “who is Obama?” are terribly difficult for computers because they require context and experience.  A tremendous amount of work went into developing algorithms that could sift a huge amount of information and come up with the most likely answer.  In the past couple of days, the computer could almost always get the right question first.  It did make critical mistakes in final jeopardy, for example, when human contestants had equal amounts of time as the computer.

There are really two parts to winning this contest.

1. Can a computer be built to answer fuzzy questions that rely on human experience?

2. Can the computer answer these questions before a highly intelligent human can?

Based on three nights of Jeopardy, the answer to the first question is a resounding Yes!   It took a lot of work, but the computer has a very good ability to not only come up with an answer, but it also has a fairly good assessment of its confidence in the answer.  During the contest a bar chart was displayed to indicate Watson’s choices and it’s calculated confidence.  Often the answer was chosen with upwards of 80% confidence.   The computer also has about 16 Terabytes of memory at its disposal – about one million books or about 4000 laptops with 4Gbytes of memory each.  Even the best human players can’t retain this much information.  But we humans do have that gift of “fuzzy logic” and uncanny recall. The real innovation is the software. That is what gave Watson the amazing ability to assess the right information and decide which solution is most credible.

Unfortunately, the answer to question #2 is uncertain.  It did appear that the contestants knew many of the answers right away, but were beaten by computer’s button push.   This did detract from the contest and obfuscated the computer’s real advantage/disadvantage vs. a human player.  The computer can push the electromechanical solenoid faster than a human can.  I see solenoids on the web that respond within 7 to 50 milliseconds (ms). I went back through the video and found that the computer actuates a solenoid within one to two video frames. That is 33 – 66 milliseconds.  Human reaction times are categorized for athletes in this excellent website on reaction times.  A good sprinter has a reaction time in the 150 ms range to a light signal.  Ken Jennings seemed a bit frustrated as he lost his lead in the last game. The fact that he had a lead at one point shows that the contestants have trained to anticipate when the light will go on. His face appeared to indicate that he knew a number of the answers Watson got right but he was scooped by the computer.

Still, this achievement by IBM should not be downplayed.  It may drive product strategies for years to come.

In only a few structured conversations with members of each customer segment you can validate the basic segmentation model.   I strongly urge you not to skip this step.   It is quite tempting for those of us who are technically or analytically inclined to want to jump to a customer survey with a large sample.   At this stage, it is much more effective understand deeply what distinguishes customer needs, and solutions options for meeting those needs.   The operating question is “Why?”  The investigation should center on revealing the customers’ deep and possibly unstated needs.  These are often buried under years of “work arounds” that customers have put in place because the novel new solution had never been available before.   The researcher can then test various product strategy ideas and begin to understand customer solution preferences.    The discovery process can be a rich experience but offers no statistical verification.  Questions like “How much will we sell?, and “What are the feature priorities?” come much later in the process.

You may be surprised when the customer does not value several features or benefits your team dreamed up to solve a particular problem.  Some segments may value them and others may not.  These unwanted features may be a goldmine for your development program!  First, everything you add to a program can increase interactions, cost and schedule.   Second, removing some of these requirements may allow you to give customers much more of what they do value.  Third, requirements differences between segments give you opportunities to differentiate your solutions and come up with a winning product strategy.

Some teams want to jump quickly to more efficient methods of learning from customers.  Focus groups may be appropriate depending on how much the product strategy deviates from existing solutions.  While a group of customers may be quite able to determine features for a new laptop, for example, they may have a very difficult time giving feedback on something like a flying car.  It is good to ask yourself:  how relevant is the experience of this customer group to the new idea?  Testing the process with two to three customers prior to full focus groups can reveal the usefulness of the approach and audience.  Focus groups tend drift to the opinions of more risk averse mainstream customers.  They tend to be a challenge when your solution is quite novel and untested.  I have often returned from focus groups and been challenged by the R&D team or management who say:  “The customers just did not understand our solution idea – we must not have presented it well”.  To this I reply:  “We had a whole hour to explain it in person.  When we launch the product, we will get at most a few seconds of customer time If they don’t understand it now, they certainly won’t understand it well enough to take a step to contact us”.   Often I find that the solution idea is so innovative that customers don’t know how to react to it.  It is simply too far out of their current thinking.  Or, the problem they have is not sufficiently important for them to stop doing what they are doing today.   These negative answers need to be carefully examined and dealt with.   There are also other difficulties with focus groups, like the tendency of some customers to dominate the conversation.  In these cases it is important to get a skilled focus group moderator who can bring out the best insights from each participant.

The process of interviewing customers can be quite labor intensive and expensive when done in person.  You are getting deep information that a skilled researcher can synthesize into deep and practical insights for the product strategy overall and for the next product release.  I have seen an excellent product strategy emerge from this step.  This is especially true in highly complex business environments where system solutions are being considered.   In that case, quantitative processes could quantify small elements of the product strategy while missing the big picture.

With this customer research in hand, the team can make a first cut of the top two or three most attractive segments to go after.  A preliminary value proposition can be created for each segment (more on that in future papers).  The next step is to refine customer segment sizing, competitive analysis, ability to deliver, simple financials, and sales motion assessment.  This information can be used with the customer research to decide the initial target segment for the first project.   Winning this segment often takes much more dedication than is anticipated and can lead to excellent growth in a new market.  While it is tempting to go after several segments at once, this should only be tried by large companies with enormous resources.

Note that in each of these steps complexity has been reduced through successive refinement so that a clear product strategy emerges with a clear value proposition to a well defined customer segment.

Summary:  Quality interviews rather than quantity contributes the most to your product strategy at this point.

Customer segments are defined as customers who have generally similar needs and desires and could be served by a similar solutions.   Solutions are the whole envelope of what the company delivers:  product, services, ordering, delivery and support.  A product strategy without these elements is not much of a product strategy and misses significant opportunities for competitive differentiation.   The segments can be considered distinct from each other for the purpose of planning solutions.    Another, often overlooked characteristic is that customers in a segment have common communication/marketing needs and may communicate with each other.  This will be invaluable later when you begin to sell solutions.

How do you know if you have a reasonable segmentation?

1. You can describe the customer in a segment with a persona.  This is a story that helps communicate who you are creating and selling a solution for.  For example:   “Sally – Soccer Mom” has two kids, and carpools to several activities per week with friends of the children.  On weekends, her husband and kids frequently go on an outing or weekend campout. (and so on).   She works outside the home and her schedule is packed from morning until night.  Her husband communtes a great distance each day.”   Although not every mother is a soccer mom exactly like Sally, there is probably a good sized group that generally fits her description.

2. The segmentation is relevant to the problem and solution you are working on.  Often marketers take variables that are easy to measure, but have little to do with understanding the customer.  The segmentation should inform your product development, and later on your sales and marketing efforts.

3. You can determine if your segments are too broad if you can identify a few much larger and capable competitors than you, who are doing well in this segment.  You need a target segment that you can dominate.

4. Ideally, the segment should have a need to communicate within itself.  Perhaps there are groups that exchange information in this segment.  If you have any hope of leveraging your sales and marketing investments, there is no better sales person than a happy customers who naturally communicate with each other.

You will have a persona for each segment you are developing for your product strategy.

The first step is to gather the team together and ask them the questions they have about the opportunities and key decisions to be made.  Your leaders can define the objectives of the study and success parameters.  Without this early definition, you may come back from the process all excited about what you learned only to find that you have “fetched the wrong rock”.  Give frequent updates to the decision makers so that they will not be surprised with the findings.  This is VITAL when the project idea was originally the brainchild of an important leader.  You may come back with the recommendation to end the program or significantly alter it.  Organize and stage your questions so that the key strategic decisions and questions are examined first.

It is highly recommended that you start the process by learning all you can from secondary sources.  Why not “stand on the shoulders of giants”, rather than reinvent the wheel?  Secondary sources are research done by others that has been published electronically or in print.  You may want to do web research or read articles, and other publications that have to do with the kind of problems customers face in their industry or application.   I would begin with search engines, then Wikipedia (mainly to get links to published papers), industry association websites, online and print magazines.  Analysts are often cited and give some insights about the market.  These point to their further services and research reports.   You may find that analyst reports are quite expensive and might be useful in certain circumstances.  I find, however that they often reflect a market that has already “left the station”.   Blogs can be a good source only if you trust the author.  As part of this process, you may find interesting people to connect with directly who have experience in this market space.   Some of them may turn into future partners or customers.

There are two important considerations.  You need to be careful about the credibility of the source, especially on the web.  Does the author mention where the data came from?  Does the author have a particular agenda such as product sales?   Similar to a journalist, you may want to validate the author’s facts.  Secondly, as you look at a new opportunity, you may want to see new ways of doing things.  If you follow the existing information too much, you may fall too much into the conventional wisdom, or existing patterns of how things are done today.   Then you are much more likely to create a me-too product strategy that will offer limited opportunities to differentiate against entrenched competitors.  Secondary research can help you understand if the segments you dreamed up have only a few customers or there is actually a suitable market (number of customers that could purchase in the long-run).   You may want to enlist a team within your company to gather the information.  It is great to get people from different functions to understand the market and customer and have meaningful dialogue about the problems.

Summary:  a product strategy that begins with secondary research stands on the shoulder of giants.

• Understand the market area.
• Define customers you want to focus on.
• Research target customer needs, wants and unstated desires.
• Examine competition.
• Determine your value proposition.
• Build the solution or product strategy roadmap.

Product development organizations often want to jump right in and develop a product without a product strategy.  Product Marketers want to promote and get the product sold.  So it is not hard to imagine that the product strategy is often developed based on a stirring together of sales force input, management team experience, and engineers who say:  ”We could do that, and it would be cool”.

A horse designed by committee

Unfortunately, this method can lead to collection of features in a long list or spreadsheet.  The target customer can easily be forgotten as the list is managed by determining how many features can be crammed into the next release.   What about the problem we’re solving for the target customer, or the benefits we will deliver? The product lifecycle meetings and checkpoints are increasingly  unpleasant with factions developing between the  teams.  Another common problem is that product managers are a bit outgunned by the development team.  During the concept phase there is a great need for customer, competitor and market research to ensure the product strategy delivers the right value to the customer.  An overwhelmed product manager will not have time to conduct the research and will be reminded that she is holding up the project.  When asked about a given feature, she may say “sure, we need that”.    Unfortunately that may delay the project and cause significant complexity without delivering results.

As the project progresses, the engineers come to marketing, hat in hand, and explain why the cost and schedule of the project will not be met unless we agree to defer some features until the next release.  Pretty soon the target customers are not likely to salute the solution that is released.   Fingers start wagging and pointing to other departments.    The sales force can’t find customers to buy the product and call up to complain about the product shortcomings.   Most often, they ask for price reductions since this is the only thing that can reasonably done at this stage.   The product manager’s task then becomes:  ”lets turn a shortcoming into a benefit”, or “lets reposition the product”, or “let’s sell this version and then promise an upgrade”.

At least in the high-tech world, where change is rapid and unpredictable, the product strategy needs to be fluid, adaptable, and yet steady in following certain principles.

I wonder how many of you out there have a story to share along these lines?

Feel free to comment on this product strategy topic.