Starbuck’s Coffee (nasd: SBUX) is experiencing a slowdown in same-store sales – specifically declines in transaction volume or “foot traffic.” This has been cause for alarm as some theorize that the business model is in jeopardy. Yet, I believe that SBUX is intact, and it’s the aggressive store expansion that has caused cannibalization. However, not positive, cannibalization can be addressed and solved, but it will entail slower growth than what has been expected.
Starbuck’s Historical Performance:
Starbucks turned out 17% sales growth for the Q1 2008, yet same-store sales increased only 1%: 2% rise in value per ticket coupled with 1% decline in transaction volume. In the US: SSS dropped 1%: 2% increase in average ticket value and 3% decline in number of transactions.
Cleary, foot traffic is declining for stores older than one year, and some are speculating that customers are defecting to McDonalds. Simply, this is not the case. Foot traffic declines are primarily a result of cannibalization from new store openings. Sales growth continues to be robust due to new store additions, yet some of that traffic is stolen from stores open longer than 13 months.
Looking back several months, Starbucks management stated on the 07Q4 conference call (11/15/07) that the 1% decline in US store traffic (# transactions) in Q4 was not due to saturation. For FY07, store traffic was basically flat. I don’t know how they could be denying the fact that their aggressive store expansions haven’t cannibalized sales from existing stores. That implies that virtually all sales generated from the 1,788 new stores opened FY07 are from customers, who hadn’t been to Starbucks in the past 13 months (FY06).
I’m really surprised that there hasn’t been a greater decline in transactions just from the simple fact that a new store addition may be more conveniently located for a regular customer than an existing store. It’s safe to assume that a portion of the new store traffic includes customers, who had been going and would still be going, to an existing location.
SBUX management is accepting the fact that cannibalization is occurring as evidenced by the announcement to dial down the number of planned new store openings.
Same Store Sales Metric Explained:
The highly popular same store sales (SSS) metric can be quite nebulous at times. Its purpose is to capture organic sales growth from existing (comparable stores) and exclude growth attributed to additions of new stores. Overall revenue growth due to new store openings can conceal falling sales at existing stores unless sales from the both segments are isolated. SSS compares sales of stores open at least a year (13m) thus excluding growth associated from an increased store count in the current period. Comps can indicate if competition is affecting sales, or if customers are spending less due to economic hardships. SSS can also uncover if the firm’s product offerings have become less relevant to consumers.
SSS can be clouded since often SSS growth is driven by new store growth, which counters the purpose and value of the SSS comparison. New store additions can affect SSS growth when the new stores require time to reach maturity (or peak revenues), which may be several months to several years. Considering a new market lacks an existing customer base, a new store will have to build a customer base by increasing awareness with promotion, advertising, and word-of-mouth. In addition, the new store will need to become familiar with local tastes and preferences in order to optimize its inventory by stocking a product assortment that best matches the surrounding community. These are just a couple factors that describe the revenue growth curve for a new store.
Individual store revenue curves widely vary due to numerous factors. The type of industry that the firm operates makes a difference. The most significant determinant is the existing customer base, specifically brand awareness and familiarity. New markets where brand penetration is absent will not have much, if any customer base. Highly visible brands in distant markets can create awareness in un-penetrated markets resulting in a more rapid build in the customer base once a new store arrives. Store additions in markets that are in close proximity to existing stores will have an existing customer base to draw from. This scenario will likely result in the quickest maturity since a portion of the customer base will be cannibalized from the other existing stores.
Retailers that are aggressively adding stores in new markets tend to have overestimated SSS figures. This is how it works: A retailer with 1000 stores opens 300 new stores, which won’t reach maturity for at least two years. The following year, the retailer now has 1300 existing stores with 300 generating 60% of maturity sales. For the year, the company adds 400 new stores bringing the total to 1700, but the SSS comparison only involves the 1300 existing at the start of the year. At year end, those 300 stores operating at 60% the year before, are now at 90% of peak store revenues. The increase in sales per store associated with those stores open longer than a year but less than time to maturity are included in SSS calculation. Thus, adding new stores will boost SSS sales if there is a considerable period between the open date and the date the customer base matures.
In general, there is a direct relationship between SSS and store count increases, especially when new stores are added in markets characterized by a small customer base that gradually increases. If possible, calculating sales per store at existing stores versus new stores can aid in uncovering situations where SSS is blurred from aggressive store expansion. If there is a wide discrepancy between the two, it’s likely that SSS is overstated. For instance, if existing stores have historically averaged 500k per store, and new stores (less 1 yr) average 250k, then SSS will grow as new stores approach peak sales of 500k. The 250k per store generated by new additions are responsible for growth associated with new stores only. Yet, for the subsequent year, the SSS metric will use 250k in comparison.
Starbuck’s Same Store Sales:
In the case of Starbuck’s, I believe that the slowdown in SSS is due to new stores having a larger initial customer base (higher sales/store) thus the comparisons are tougher. Stores coming online generating sales/store closer to peak sales and maturing more quickly will result in higher new store sales growth and lower SSS growth in the subsequent period.
A logical assumption is that during SBUX expansion into new geographic markets, new stores gradually grew traffic as brand exposure increased with time. In contrast, stores opened in already well penetrated markets require less effort drawing traffic since many consumers have already experienced Starbucks’s offerings. SBUX current store footprint virtually blankets the US, therefore a significant portion of store additions will be in locations in the vicinity of existing SBUX stores. SBUX is highly visible, well-established brand that reduces the local effort needed to inform and attract a customer base because of the large percentage of local consumers is already familiar.
Many new Starbuck’ stores provide a more convenient location for some of its current customers who are regular visitors at more distantly located stores. Thus, SBUX customers switch to the new location. It’s also likely that some customers opt for a new location to avoid the high traffic stores. Yet, it’s hard to argue that sales generated at new locations do not come at the expense of existing stores. How much? It’s hard to say without conducting extensive market research and customer surveys.
I have little doubt that cannibalization is responsible for the decline in SSS transaction growth. Traffic migrating from an existing store to a new store, causes SSS revenue to fall, but total revenue is unaffected because of the offsetting increase in new store sales growth. The overall effect is negative because the same level of revenue is spilt between two stores instead of one, thus lowering margins and return on invested capital. Yet, if new stores increase an existing customers’ visits/yr and allow existing stores to capture new customers as well, then growth could outweigh cannibalization.
According to my calculations, for the past couple years, it appears that revenues at new stores are closer to revenue/store at existing locations than in previous years. This suggests more recent store openings have a larger initial customer base.
Turning Cannibalization into Positive Growth:
It’s possible that this type of cannibalization can actually end up being positive for SBUX if several factors occur.
1) Transactions/Customer are higher at new locations than what was lost from existing stores.
2) Costs for new store openings are lower.
3) Relieve capacity pressures at high volume locations
It’s positive if a customer who averaged one visit/month at an SBUX is now averaging five visits/month at the new location. SSS will fall because store visits have fallen from 12/yr to zero. But, new store sales will increase by an amount larger (60/yr) than that lost in SSS (12/yr).
If SBUX can migrate regular, high frequency visitors from a high volume store to a new store, then that will free up capacity to attract new and low frequency customers at the high visibility locations. Premium locations serve as a marketing tool as well. The SBUX brand is exposed to the large sums of people passing by daily. Since Starbucks historically has not aggressive engaged in advertising, the stores serve as its primary marketing effort.
High traffic at a SBUX can be both beneficial and disadvantageous for attracting consumers with minimal brand experience. High volume indicates that the offerings are very liked and demanded, and provides a vote of confidence for those unsure due to lack of brand experience. Yet, some consumers may reluctant because they are unwilling to deal with long lines.
Consumers willing to engage crowded locations are more likely to be regular visitors who perceive SBUX product value to exceed the cost of waiting. The cost of a premium location is worthwhile when the additional expense for visibility attracts new customers or increases visits of low frequency customers. The additional expense provides less benefit if the highly visible locations are packed with regular customer who “crowd out” curious, potential customers.
The ideal scenario: high frequency visitors go to convenient, but less visible (lower store cost) locations thereby freeing up service capacity at the premium locations. This will allow the highly visible, premium locations to attract more new / seldom customers.
Anecdotal Evidence:
To give an example, I will describe my recent Starbuck’s visiting pattern. Of course, this is just my experience and may not accurately represent the norm. Yet, I can envision that a portion of SBUX customers share a similar pattern. I think it’s at least worth mentioning and contemplating.
The SBUX store closest to my house was a premium location in mid-town. During morning rush hour, the drive-thru line spills all the way out into the street. Occasionally, I couldn’t get in line because I would be blocking traffic at the intersection so I would have to keep moving. At this time, I probably visit SBUX once every two months (6 visits/yr). This distance from my house and the large crowds at SBUX dampened my visit frequency. Several months ago, a small SBUX opened close to my house downtown. With much smaller lines and the convenient location, I now go to Starbucks at least twice a month (24 visits/yr) and usually spend more too.
Now- my visits have increased, my average ticket has increased, and the new location is likely less expensive to open and operate. In addition, since I no longer visit the high traffic SBUX, there will be smaller lines to capture customers who may pass-up due to the high volume.
I have asked several acquaintances what would influence them to visit SBUX more frequently, and the top responses I received were convenient location and shorter lines. A few essentially mentioned they don’t think about SBUX until they see a store, but usually the road traffic and store parking lot appear to be a hassle.
My experience is just an anecdotal scenario that has little value unless we could ascertain the exact percentage of consumers who share my experience. Yet, I am certain that some do, and at minimum, it’s a worthy scenario to a least consider.
Conclusion:
Starbucks needs to manage its store expansion strategy more effectively to reduce the negative effects of cannibalization. Situations where cannibalization is unavoidable, SBUX needs to find methods to limit the negative impact to achieve scenarios I mentioned above. Essentially, this involves selecting locations that can be supported by an existing customer base coupled with potential to attract new customers. It appears that SBUX is beginning to understand the issue from the recent announcement to slow store expansion as well as closing underperforming locations. Management plans to be meticulous with respect to store additions. A scaled back expansion plan will translate into slower sales and earnings growth, hence price multiple. It appears that this information is reflected in the stock price as the multiple has already been compressing for some time.
I believe that sales and earnings growth, albeit a slower rate, will still be above average and persist for a considerable length of time. Shares are fairly valued currently, yet as a long-term hold, SBUX is relatively attractive.
Disclosure: No position in SBUX
Starbuck’s Historical Performance:
Starbucks turned out 17% sales growth for the Q1 2008, yet same-store sales increased only 1%: 2% rise in value per ticket coupled with 1% decline in transaction volume. In the US: SSS dropped 1%: 2% increase in average ticket value and 3% decline in number of transactions.
Cleary, foot traffic is declining for stores older than one year, and some are speculating that customers are defecting to McDonalds. Simply, this is not the case. Foot traffic declines are primarily a result of cannibalization from new store openings. Sales growth continues to be robust due to new store additions, yet some of that traffic is stolen from stores open longer than 13 months.
Looking back several months, Starbucks management stated on the 07Q4 conference call (11/15/07) that the 1% decline in US store traffic (# transactions) in Q4 was not due to saturation. For FY07, store traffic was basically flat. I don’t know how they could be denying the fact that their aggressive store expansions haven’t cannibalized sales from existing stores. That implies that virtually all sales generated from the 1,788 new stores opened FY07 are from customers, who hadn’t been to Starbucks in the past 13 months (FY06).
I’m really surprised that there hasn’t been a greater decline in transactions just from the simple fact that a new store addition may be more conveniently located for a regular customer than an existing store. It’s safe to assume that a portion of the new store traffic includes customers, who had been going and would still be going, to an existing location.
SBUX management is accepting the fact that cannibalization is occurring as evidenced by the announcement to dial down the number of planned new store openings.
Same Store Sales Metric Explained:
The highly popular same store sales (SSS) metric can be quite nebulous at times. Its purpose is to capture organic sales growth from existing (comparable stores) and exclude growth attributed to additions of new stores. Overall revenue growth due to new store openings can conceal falling sales at existing stores unless sales from the both segments are isolated. SSS compares sales of stores open at least a year (13m) thus excluding growth associated from an increased store count in the current period. Comps can indicate if competition is affecting sales, or if customers are spending less due to economic hardships. SSS can also uncover if the firm’s product offerings have become less relevant to consumers.
SSS can be clouded since often SSS growth is driven by new store growth, which counters the purpose and value of the SSS comparison. New store additions can affect SSS growth when the new stores require time to reach maturity (or peak revenues), which may be several months to several years. Considering a new market lacks an existing customer base, a new store will have to build a customer base by increasing awareness with promotion, advertising, and word-of-mouth. In addition, the new store will need to become familiar with local tastes and preferences in order to optimize its inventory by stocking a product assortment that best matches the surrounding community. These are just a couple factors that describe the revenue growth curve for a new store.
Individual store revenue curves widely vary due to numerous factors. The type of industry that the firm operates makes a difference. The most significant determinant is the existing customer base, specifically brand awareness and familiarity. New markets where brand penetration is absent will not have much, if any customer base. Highly visible brands in distant markets can create awareness in un-penetrated markets resulting in a more rapid build in the customer base once a new store arrives. Store additions in markets that are in close proximity to existing stores will have an existing customer base to draw from. This scenario will likely result in the quickest maturity since a portion of the customer base will be cannibalized from the other existing stores.
Retailers that are aggressively adding stores in new markets tend to have overestimated SSS figures. This is how it works: A retailer with 1000 stores opens 300 new stores, which won’t reach maturity for at least two years. The following year, the retailer now has 1300 existing stores with 300 generating 60% of maturity sales. For the year, the company adds 400 new stores bringing the total to 1700, but the SSS comparison only involves the 1300 existing at the start of the year. At year end, those 300 stores operating at 60% the year before, are now at 90% of peak store revenues. The increase in sales per store associated with those stores open longer than a year but less than time to maturity are included in SSS calculation. Thus, adding new stores will boost SSS sales if there is a considerable period between the open date and the date the customer base matures.
In general, there is a direct relationship between SSS and store count increases, especially when new stores are added in markets characterized by a small customer base that gradually increases. If possible, calculating sales per store at existing stores versus new stores can aid in uncovering situations where SSS is blurred from aggressive store expansion. If there is a wide discrepancy between the two, it’s likely that SSS is overstated. For instance, if existing stores have historically averaged 500k per store, and new stores (less 1 yr) average 250k, then SSS will grow as new stores approach peak sales of 500k. The 250k per store generated by new additions are responsible for growth associated with new stores only. Yet, for the subsequent year, the SSS metric will use 250k in comparison.
Starbuck’s Same Store Sales:
In the case of Starbuck’s, I believe that the slowdown in SSS is due to new stores having a larger initial customer base (higher sales/store) thus the comparisons are tougher. Stores coming online generating sales/store closer to peak sales and maturing more quickly will result in higher new store sales growth and lower SSS growth in the subsequent period.
A logical assumption is that during SBUX expansion into new geographic markets, new stores gradually grew traffic as brand exposure increased with time. In contrast, stores opened in already well penetrated markets require less effort drawing traffic since many consumers have already experienced Starbucks’s offerings. SBUX current store footprint virtually blankets the US, therefore a significant portion of store additions will be in locations in the vicinity of existing SBUX stores. SBUX is highly visible, well-established brand that reduces the local effort needed to inform and attract a customer base because of the large percentage of local consumers is already familiar.
Many new Starbuck’ stores provide a more convenient location for some of its current customers who are regular visitors at more distantly located stores. Thus, SBUX customers switch to the new location. It’s also likely that some customers opt for a new location to avoid the high traffic stores. Yet, it’s hard to argue that sales generated at new locations do not come at the expense of existing stores. How much? It’s hard to say without conducting extensive market research and customer surveys.
I have little doubt that cannibalization is responsible for the decline in SSS transaction growth. Traffic migrating from an existing store to a new store, causes SSS revenue to fall, but total revenue is unaffected because of the offsetting increase in new store sales growth. The overall effect is negative because the same level of revenue is spilt between two stores instead of one, thus lowering margins and return on invested capital. Yet, if new stores increase an existing customers’ visits/yr and allow existing stores to capture new customers as well, then growth could outweigh cannibalization.
According to my calculations, for the past couple years, it appears that revenues at new stores are closer to revenue/store at existing locations than in previous years. This suggests more recent store openings have a larger initial customer base.
Turning Cannibalization into Positive Growth:
It’s possible that this type of cannibalization can actually end up being positive for SBUX if several factors occur.
1) Transactions/Customer are higher at new locations than what was lost from existing stores.
2) Costs for new store openings are lower.
3) Relieve capacity pressures at high volume locations
It’s positive if a customer who averaged one visit/month at an SBUX is now averaging five visits/month at the new location. SSS will fall because store visits have fallen from 12/yr to zero. But, new store sales will increase by an amount larger (60/yr) than that lost in SSS (12/yr).
If SBUX can migrate regular, high frequency visitors from a high volume store to a new store, then that will free up capacity to attract new and low frequency customers at the high visibility locations. Premium locations serve as a marketing tool as well. The SBUX brand is exposed to the large sums of people passing by daily. Since Starbucks historically has not aggressive engaged in advertising, the stores serve as its primary marketing effort.
High traffic at a SBUX can be both beneficial and disadvantageous for attracting consumers with minimal brand experience. High volume indicates that the offerings are very liked and demanded, and provides a vote of confidence for those unsure due to lack of brand experience. Yet, some consumers may reluctant because they are unwilling to deal with long lines.
Consumers willing to engage crowded locations are more likely to be regular visitors who perceive SBUX product value to exceed the cost of waiting. The cost of a premium location is worthwhile when the additional expense for visibility attracts new customers or increases visits of low frequency customers. The additional expense provides less benefit if the highly visible locations are packed with regular customer who “crowd out” curious, potential customers.
The ideal scenario: high frequency visitors go to convenient, but less visible (lower store cost) locations thereby freeing up service capacity at the premium locations. This will allow the highly visible, premium locations to attract more new / seldom customers.
Anecdotal Evidence:
To give an example, I will describe my recent Starbuck’s visiting pattern. Of course, this is just my experience and may not accurately represent the norm. Yet, I can envision that a portion of SBUX customers share a similar pattern. I think it’s at least worth mentioning and contemplating.
The SBUX store closest to my house was a premium location in mid-town. During morning rush hour, the drive-thru line spills all the way out into the street. Occasionally, I couldn’t get in line because I would be blocking traffic at the intersection so I would have to keep moving. At this time, I probably visit SBUX once every two months (6 visits/yr). This distance from my house and the large crowds at SBUX dampened my visit frequency. Several months ago, a small SBUX opened close to my house downtown. With much smaller lines and the convenient location, I now go to Starbucks at least twice a month (24 visits/yr) and usually spend more too.
Now- my visits have increased, my average ticket has increased, and the new location is likely less expensive to open and operate. In addition, since I no longer visit the high traffic SBUX, there will be smaller lines to capture customers who may pass-up due to the high volume.
I have asked several acquaintances what would influence them to visit SBUX more frequently, and the top responses I received were convenient location and shorter lines. A few essentially mentioned they don’t think about SBUX until they see a store, but usually the road traffic and store parking lot appear to be a hassle.
My experience is just an anecdotal scenario that has little value unless we could ascertain the exact percentage of consumers who share my experience. Yet, I am certain that some do, and at minimum, it’s a worthy scenario to a least consider.
Conclusion:
Starbucks needs to manage its store expansion strategy more effectively to reduce the negative effects of cannibalization. Situations where cannibalization is unavoidable, SBUX needs to find methods to limit the negative impact to achieve scenarios I mentioned above. Essentially, this involves selecting locations that can be supported by an existing customer base coupled with potential to attract new customers. It appears that SBUX is beginning to understand the issue from the recent announcement to slow store expansion as well as closing underperforming locations. Management plans to be meticulous with respect to store additions. A scaled back expansion plan will translate into slower sales and earnings growth, hence price multiple. It appears that this information is reflected in the stock price as the multiple has already been compressing for some time.
I believe that sales and earnings growth, albeit a slower rate, will still be above average and persist for a considerable length of time. Shares are fairly valued currently, yet as a long-term hold, SBUX is relatively attractive.
Disclosure: No position in SBUX
