Granularization is what I call the method I utilize to deconstruct and question the fundamental assumptions of a strategy. Here are the steps:
1) Break down the product's value proposition and identify how it interrelates with the company's overall strategy.
2) Determine the most basic variables that would confirm the assumptions that underpin the strategy.
3) Once this determination is made, then discover the company's justification for the strategy. Use this information to further refine your results on top of the basic assumptions that you analyze.
While granularization's use is limited and subject to confirmation bias, I have found it very useful for evaluating a strategy from a fresh perspective.
I followed Zipcar's progress since before its IPO in early 2011. Innovative, hip, a bit edgy: what's not to like? Only when I bored down into the details did it seem off.
1) Value proposition: car sharing (many "members" paying to share the same cars). Zipcar maintains a minimal physical presence to occasionally service the vehicles. Systems and call centers handle the rest. This could, in theory and when operating in large cities, hold costs to less than traditional rental car companies, but with more locations.
2) Basic Assumptions:
The Data:
A) Density:
Cities with more than 4,000 housing units per square mile are uncommon in the US. All cities with more than 10,000 housing units per square mile are located in the greater NYC area. Accordingly, one would expect atypical results from this region. To give an idea on scale, the 200-250 range contains cities such as Las Vegas, Norfolk, and Grand Rapids; these are all moderately sized cities, but they have relatively few buildings above five floors.
B) Income:
The above graph shows all cities with a density of over 4,000 households per square mile (excluding the NYC region cities above 10,000 for truncation purposes). Each of these cities is then grouped into major metropolitan areas.
Looking at the data available, I believe there are two key points: (1) there are very few dense cities in the US in which Zipcar can optimally operate, and (2) as Zipcar expands into cities farther from the high-density/high-income quadrant, one cannot expect similarly good results from these cities.
3) Company Reasoning (at least the publicly available information): Soon after Zipcar's IPO, I reviewed a number of explanations and projections given by management for the company's progress and future (each is paraphrased).
A) In time Zipcar's newer cities will mature and obtain results similar to NYC, Boston, etc. Zipcar initially launched its operations in most of its ideal markets (NYC, Boston, DC). Because Zipcar's mature markets are outliers in terms of density and income, replication of NYC per-car-volume and margins in Cleveland, Denver, or Austin can't happen.
B) Specific sub-market segments (e.g., university campuses) can be targeted and profitable. Zipcar lacks the infrastructure to execute on this very well, and this opportunity does not appear significant enough to drive the aggressive forecasts. It is a possibility though, and specific market research and pilots should bear out this further refinement of the data.
C) Primary cities will grow at a fast pace. This is reasonable and should be very easy to show with data. Continued fast-pace growth can be demonstrated with gross hours rented by month in that city (not just members enrolled).
D) Opportunities outside the US will continue the growth trajectory. Certainly possible, but the execution would be quite complicated, and up until the IPO Zipcar had not demonstrated complex international expansion as a core competency.
The Result: Zipcar IPO'd at $18 and popped to almost $30 based on buzz and aggressive projections. It then steadily fell over the next couple years to $7 as Zipcar continued to essentially break even without achieving its aggressive revenue targets and could not demonstrate success in the four projections shown above. Zipcar was eventually bought out by Avis with a premium that amounted to approximately 25% less than Zipcar's then 52-week high.
Granularization is especially effective at breaking apart "but this time it's different" arguments by beginning with a straightforward "blunt instrument" and then taking into account company justifications after the fact.
From Bloomberg:
Notes on Data: In order to get a consistent and complete data set with minimal judgement calls, income was pulled on a metropolitan area basis from FRED, while housing density was pulled from the census (GCT-PH1-Geography-United States: Population, Housing Units, Area, and Density: 2010). While better (public) data sets for an exercise such as this do exist, too much discretion on data classification than I am generally comfortable with would need to be used in order to generate an "apples-to-apples" comparison. Additionally, housing density by metropolitan area was not used because the results can vary considerably depending on how metropolitan area is defined. Accordingly, the more specific city-level unit of data for housing density was used.
1) Break down the product's value proposition and identify how it interrelates with the company's overall strategy.
2) Determine the most basic variables that would confirm the assumptions that underpin the strategy.
3) Once this determination is made, then discover the company's justification for the strategy. Use this information to further refine your results on top of the basic assumptions that you analyze.
While granularization's use is limited and subject to confirmation bias, I have found it very useful for evaluating a strategy from a fresh perspective.
I followed Zipcar's progress since before its IPO in early 2011. Innovative, hip, a bit edgy: what's not to like? Only when I bored down into the details did it seem off.
1) Value proposition: car sharing (many "members" paying to share the same cars). Zipcar maintains a minimal physical presence to occasionally service the vehicles. Systems and call centers handle the rest. This could, in theory and when operating in large cities, hold costs to less than traditional rental car companies, but with more locations.
2) Basic Assumptions:
A) Density. Zipcar needs a sufficient number of people within walking distance of the vehicle. Because of Zipcar's minimalist physical presence, the company isn't structured like a hub and spoke rental car company with the infrastructure to support scattered locations. The cars have to be where the people are, and the economies of scale need to be such that the cars can be serviced and cleaned on a regular basis by relatively few people.
B) Money. Providing a $15k-$25k piece of insured equipment isn't cheap, even with innovative technology and a creative business model. A $30 fee to transport items from a grocery or supply store and back isn't attractive for most of the population. There are far more substitutes in this segment than most people initially imagine (delivery service, taxis, friends, family, bikes, walking, rental cars, public transit, or simply going without and pursuing more local options).
The Data:
A) Density:
Cities with more than 4,000 housing units per square mile are uncommon in the US. All cities with more than 10,000 housing units per square mile are located in the greater NYC area. Accordingly, one would expect atypical results from this region. To give an idea on scale, the 200-250 range contains cities such as Las Vegas, Norfolk, and Grand Rapids; these are all moderately sized cities, but they have relatively few buildings above five floors.
B) Income:
The above graph shows all cities with a density of over 4,000 households per square mile (excluding the NYC region cities above 10,000 for truncation purposes). Each of these cities is then grouped into major metropolitan areas.
Looking at the data available, I believe there are two key points: (1) there are very few dense cities in the US in which Zipcar can optimally operate, and (2) as Zipcar expands into cities farther from the high-density/high-income quadrant, one cannot expect similarly good results from these cities.
3) Company Reasoning (at least the publicly available information): Soon after Zipcar's IPO, I reviewed a number of explanations and projections given by management for the company's progress and future (each is paraphrased).
A) In time Zipcar's newer cities will mature and obtain results similar to NYC, Boston, etc. Zipcar initially launched its operations in most of its ideal markets (NYC, Boston, DC). Because Zipcar's mature markets are outliers in terms of density and income, replication of NYC per-car-volume and margins in Cleveland, Denver, or Austin can't happen.
B) Specific sub-market segments (e.g., university campuses) can be targeted and profitable. Zipcar lacks the infrastructure to execute on this very well, and this opportunity does not appear significant enough to drive the aggressive forecasts. It is a possibility though, and specific market research and pilots should bear out this further refinement of the data.
C) Primary cities will grow at a fast pace. This is reasonable and should be very easy to show with data. Continued fast-pace growth can be demonstrated with gross hours rented by month in that city (not just members enrolled).
D) Opportunities outside the US will continue the growth trajectory. Certainly possible, but the execution would be quite complicated, and up until the IPO Zipcar had not demonstrated complex international expansion as a core competency.
The Result: Zipcar IPO'd at $18 and popped to almost $30 based on buzz and aggressive projections. It then steadily fell over the next couple years to $7 as Zipcar continued to essentially break even without achieving its aggressive revenue targets and could not demonstrate success in the four projections shown above. Zipcar was eventually bought out by Avis with a premium that amounted to approximately 25% less than Zipcar's then 52-week high.
Granularization is especially effective at breaking apart "but this time it's different" arguments by beginning with a straightforward "blunt instrument" and then taking into account company justifications after the fact.
From Bloomberg:
Notes on Data: In order to get a consistent and complete data set with minimal judgement calls, income was pulled on a metropolitan area basis from FRED, while housing density was pulled from the census (GCT-PH1-Geography-United States: Population, Housing Units, Area, and Density: 2010). While better (public) data sets for an exercise such as this do exist, too much discretion on data classification than I am generally comfortable with would need to be used in order to generate an "apples-to-apples" comparison. Additionally, housing density by metropolitan area was not used because the results can vary considerably depending on how metropolitan area is defined. Accordingly, the more specific city-level unit of data for housing density was used.
No comments:
Post a Comment