One of our first clients in the Automotive Aftermarket was a manufacturer of replacement parts. They had developed their own in-house methodology for generating inventory stocking recommendations for their customers. They didn’t need us to reinvent the wheel. But, they had problems applying their methodology efficiently.
They had built their system in Microsoft Excel and it was a fairly complicated process to generate SKU-level recommendations. It required a lot of copy-paste to get the right data in the right spreadsheet cells. And the manual data input process led to frequent mistakes and frustrations. Because of this, they had 2 full-time analysts who spent almost all their time waist-deep in the spreadsheet trying to keep up with requests for data-driven recommendations.
We suggested that we take their approach and spreadsheet and convert it into a web application. This had several benefits and a rapid ROI:
In my last post we looked at the population of Honda Accords in New York. We calculated a few descriptive statistics and examined a few data distributions. This helped us get to know the Honda Accord population in New York but there’s a lot more we can do to understand this potential population of aftermarket part customers.
We now know the size of the Accord population but what about the rate of Accord ownership? Where are Accords popular or unpopular? Of course there are more Accords in the New York City metropolitan area where there are millions of people. But where do people own Accords at higher or lower rates? And how do these ownership rates compare with similar vehicles? This may help us avoid having too many or too few parts on the shelf in certain locations.
This simple question is at the core of what VIO data are all about.
Imagine the entrepreneur who has invented an innovative aftermarket replacement part. The part might fit dozens of different models but, for the sake of a simple illustration, let’s say it only fits the Honda Accord. So, finding out the vehicle population is the same as finding out the size of the market for this new part. And this is the denominator in the equation to calculate market share. Pretty important number, explaining why companies are willing to spend money on VIO.
According to data we downloaded and processed from data.ny.gov, the answer to the question is there are approximately 300,000 Honda Accords in New York state.
Hey, why not give the exact number? Because all VIO data are estimates.
Even if we retrieve the data today, this minute, it would be just a static snapshot of registration records. New vehicle sales in the past month won’t show up. Recently purchased vehicles registered in a neighboring state won’t show up. Recently scrapped vehicles will still be counted. Occasionally people drive to another state and need a repair. Heck, some people register their vehicle in another state…for tax reasons or whatever. Sometimes vehicle registrations are clustered in a corporate or government office location even though the vehicles themselves are spread around the state or the country. It doesn’t matter how hard you try, the best count you can get is an estimate. If you’re thinking you’ll gain significant advantage from obtaining higher levels of precision you may be missing the mathematical forest for the trees.
In a recent article the Economist declared that Data has surpassed Oil as the world’s most valuable commodity. Ironically, this value transition is also occurring in the automotive industry where oil still fuels the cars on the road but key data, such as vehicles in operation (VIO), now fuel decisions faced by automotive parts suppliers, distributors and retailers who are increasingly competing on analytics and supply chain efficiency.
For those not familiar, VIO is the name used to describe the census of vehicle ownership registrations which include details such as year, make, model and other attributes. Most large companies in the $300B+ Automotive Aftermarket market license these data in one of two flavors, “Experian” or “Polk”. I don’t know who is winning the market share battle between the two but I do know the Aftermarket is paying a hefty premium due to inadequate competition. Buyers can choose the geographic unit of reporting (national, state, county, ZIP, census tract or perhaps even block group) with prices going up with each increase in geographic resolution. I don’t have details on pricing but, from what I can gather, you’ll pay around $20-30k for County level, maybe $40-50k for ZIP and possibly over $100k for data at census tract or block group level. In exchange for this hefty price tag you get a pile of DVDs or a monster CSV file to download, and not much more.
Originally, R.L. Polk & Company had, essentially, a monopoly on the VIO data licensing business and they used their profit margins intelligently, acquiring Carfax in 1999. Then, around 10 years ago Experian (yes, the same Experian who recently exposed the private data of 123 million American consumers) entered the market. I assume this duopoly improved the competitive landscape a bit for VIO buyers but I think there’s still plenty of room on the field. In 2013, IHS-Markit acquired R.L. Polk & Co for the tidy sum of $1.4 Billion. Having been through a couple of acquisitions myself I suspect some of the Polk brain trust has cashed-in most or all of their Earnout and moved on, either literally or figuratively.
An important piece of the automotive aftermarket category management puzzle involves an understanding of your category’s replacement rates. Replacement rates, which are also referred to as repair rates or failure rates, are essentially an estimate of the likelihood that a vehicle will need to a replacement part due to failure or normal wear and tear.
So, how should replacement rates be calculated?
Well, it starts with determining an appropriate numerator and denominator. The denominator should represent an estimate of the total population of vehicles. The numerator should represent an estimate of the total number of vehicles that required a particular part replacement.
When it comes to forecasting the future I like to think of two quotes from one of the smartest people I’ve ever worked with (the quote may not be precise but hopefully you’ll get the idea).
1. Forecasts are always wrong.
2. Forecasts with longer time horizons are always worse.
David Simchi-Levi brilliant MIT Professor and my former boss at LogicTools (now part of IBM), told me this in person and I’m pretty sure he’s expressed the same idea in one or more of his many now-famous supply chain related publications.
I thought of these quotes immediately when I read a special report published in November by the Automotive Aftermarket Supply Association titled, “Don’t Discount Miles Driven in Long Term Forecasts”.
A severe outbreak of cholera was ravaging the Soho neighborhood of London during the summer of 1854. While city leaders and health officials struggled to determine the source of the outbreak, a physician named John Snow had a groundbreaking idea: he used a map of the neighborhood to plot cases of cholera. Ultimately, his method not only revealed the source of the outbreak, but also provided new information about the nature of the disease itself.
Data mapping photograph
Dr. Snow’s work is often considered to be the founding event in modern epidemiology. For me, it’s also a great illustration of the power of a simple map and its ability to illuminate patterns in data. Much like the citizens of London, business analysts throughout the world struggle to understand and diagnose pain points in their businesses. Why are we having trouble keeping SKU# 123 in stock in Morgantown, West Virginia? How can we minimize the cost of getting parts to installers in Wyoming? How can we avoid costly product returns? Sound familiar?
Until recently the Automotive Aftermarket was provided data from key channel distributors indicating monthly sales activity and market share for various vehicle part categories. At the beginning of 2012 the consortium of companies that provided these data collapsed. Since then, parts suppliers and others in the Aftermarket have been searching for a new source of data to fill the void and this very issue is being discussed by industry representatives at the AAIA Fall Leadership Days conference in San Francisco.
In this post I propose a methodology for estimating market size and discuss how this estimate of total demand can form the basis for replacing, and perhaps improving upon, the market data previously provided by NPD.