Rumon is now three. Yesterday we went grocery shopping at Ralph’s. As always, he darted straight to the cookies section and picked up his favorite pecan butter shortbread variety. He has liked these cookies for a long time now. He even likes the box that they come in – a shiny bright blue, with yellow stars strewn at places . He is absolutely set in his likes and dislikes now! So yesterday, I noticed a bright purple box of pecan butter chocolate chip cookies a couple shelves away. Purple, pecan butter, chocolate chips – all traits ticked the right boxes in Little Mister’s like-book. “This is bound to be a hit”, I confidently declared to my husband.
Nope. Rumon did not care for my new find at all.
Sometimes I tend to forget what a difficult customer my son is. How short-lived his likes and dislikes can be, and how difficult to implement can a coercion strategy be. And when it comes to getting him to try something new, how wrong my predictions can be!
If I was a product manager introducing a new product to my market, a similar consequence would have indicated ‘faulty product forecasting’. Product Forecasting involves predicting the market for your product. For a product manager, it is one of the most important activities and formidable skills in his repertoire. What makes it a difficult skill to acquire is the tacit market knowledge that invariably comes with experience in product management. Let us get back to our analogy from parenting a toddler. Given the fact that I have more hands-on experience in dealing with Rumon’s everyday life than my husband has, who do you think has more tacit knowledge and therefore more probability of success at predicting his likes? Me, right? Despite this, what do you make of my failed attempt at predicting the ‘hit’ that I thought the chocolate chip cookies were going to be? The answer lies in understanding the difference between new product forecasting and existing product forecasting.
In mature markets, forecasting the performance of existing products is relatively straightforward. The product manager usually builds a prediction algorithm for each product line, and keys in the values for the variables in the algorithm to arrive at a (usually robust) number predicting the product performance (read sales) for the next period. This function is usually a factor of historical sales for the product, warranty and costs of repair/maintenance data, product issues and overhauls, predicted development costs (if the product line is due for a refurbish or redevelopment), and a mix of other macro economic indicators such as inflation. To predict Rumon’s ‘existing’ pecan butter shortbread cookie intake for the next month, for instance, I need to take into consideration facts such as: how many of his these cookies did he consume last month, how many did I end up consuming myself to prevent wastages (there are other implications to this angle, but let’s save that discussion for another relevant day ;-)), has this cookie intake ever had ill effects on his tummy, and how does this cookie purchase fit into our overall monthly grocery budget for next month?
Predicting new product performance is much more complicated than that. Much of the complication arises from the fact that there is little or no historical data to base your predictions on. So, most of the time, the product manager needs to talk to customers and collect their needs. The biggest challenge with this approach is that, more often than not, the customers do not know all the possibilities and are providing their opinion based on knowledge that is more restricted than the product manager’s himself. Then there is the ultimate product management quandary that has been so expertly put to words by Steve Jobs:
“You can’t just ask customers what they want and then try to give it to them. By the time you get it built. they’ll want something new”
Of course, working against this paradox and launching successful products is just what makes the product manager’s position unique. He needs the foresight to preempt what the customer would need next. He needs the vision to translate today’s vague requirements, as stated by the customer, into tangible product feature information for the next launch. A successful product manager knows that disregarding the customer’s voice could land a deathblow to his product launch plans. There are several examples of catastrophic product launches attributed to incorrect translation of the customer’s needs. One of my favorites though, is that of the failure of Ford Edsel in 1957 leading to a whopping loss of $ 350 millions by Ford. Of course, product management was not a distinctly developed role in any organization that far back in the 20th century. That fact, though, does not assuage the hurt from Ford’s wound.
For predicting the performance of new products, product managers have tried to take a qualitative approach (as opposed to the curt mathematical approach for existing products). Assumptions are laid out that provide transparency, and contingency plans are chalked out for different assumptions regarding the new product’s success or failure. What price will be right, what development cost will be incurred, what will be the cost of producing a unit: various scenarios for each of these considerations are constructed. Most product managers would argue that the accuracy of these assumptions is often determined by how well have the customer’s needs been translated.
Of course, most product managers would also hope that not many of their customers have as unpredictable buying behavior as my 3 year old’s!