Editor's Note: While this article may not be applicable or of interest to small wineries, the reality of selling wine to the large chain accounts is becoming increasingly common as well as complicated. While this topic applies only to a narrow band of our readers, we feel the trend of chain sales and the dominance of the chain markets make this an important issue for us to address. Please note that the author is a consultant whose business is based on this topic and he includes his model for data agreements. We provide this as information, not as an endorsement of one product over another.

The chains have power. There seems to be an endless line of wineries wanting shelf or wine list space, each hoping for greater sales. Agreements can help position your winery to increased sales.

One effective way of doing this is to secure an agreement at the chain level that influences individual stores to feature a product. Obtaining such an agreement often involves considerable expense to the supplier. This expense can take the form of a time commitment by the sales force, pricing discounts and other promotional considerations. If thought of in terms of an investment, these initial expenses must offer a return in the form of increased sales.

Once secured, the performance of such an agreement can be measured by the actual sales generated at the store level. These sales should be attributed to the agreement only to the extent that they are greater than what would have occurred independent of any agreement being in place. To say that an agreement was successful for the supplier, the profit generated by the incremental sales that can be attributed to the agreement must cover the associated expenses and generate an acceptable rate of return.

Clearly, the performance of the agreement with the chain is tied to the performance of the featured product at each store. The performance at each store is dependent on several factors:

• The product is readily available for consumer purchase

• The store's personnel are knowledgeable with respect to the product and are willing to direct consumers toward the product

• Promotional components of the agreement such as discounted pricing are consistently passed along to consumers

Ideally, the agreement with the chain would be sufficient to ensure that terms of the agreement would be implemented at each store. Real world experience shows this not always to be the case. This situation is further complicated in the wine industry where multiple distributors are involved as necessitated by the three-tier distribution system.

Performance of the Agreement

When a chain agrees to feature products, it is agreeing to incur certain expenses. For on-premise chains, these expenses can take the form of printing revised menus and wine lists. Off-premise chains also incur expenses--in their case related to design and purchase of print advertising, and the redesign of displays. Both types of chains may have expenses relating to the training of their employees on the featured products.

Because of the expenses involved, chains have an expectation of service from both the supplier and the distributors so that the launch of a featured product is implemented without incident. The following is a partial list of problems that can occur at the initial launch of a featured product:

• The supplier does not have the featured product available for shipment

• Distributors have not placed orders for the product

• Individual stores have not placed orders with their distributor for the product

• Shipping errors or delays result in the product being unavailable at an individual store

The end result of any of these breakdowns in the supply chain is that consumers seeing the product in advertising or on a wine list or menu will be disappointed to find the product is unavailable. Consumers, with alternate products readily available, will make a different selection. Employees at the stores, being the bearers of bad news to the consumer, will develop a negative perception of the brand. Management of the chain, given enough instances of product unavailability, will seek out alternative suppliers for future agreements.

Regardless of where the breakdown in the supply chain occurred, the resulting negative perception by the consumer, store level employees, and the chain's management will mostly be directed towards the brand and supplier. Thus the supplier has the greatest incentive to ensure that all aspects of the launch of a featured product are effectively managed.

The most useful resource available to a supplier to manage the launch of the product is its own local sales force. To fulfill this role, the suppliers' representatives must be aware of the agreement and must monitor the performance of both the distributors and the individual stores. This monitoring may take the form of tracking inventory levels, adherence to pricing and involvement in resolving problems as they arise.

Distributors have an incentive to ensure the success of these types of agreements since the initial expenses are paid by the supplier and the resulting sales require little more than order taking on the part of the distributor. One factor that can limit the success of a distributor is simply awareness. A distributor cannot secure inventory for a featured product launch they know nothing about.

In the case of chains whose stores are spread across a wide geographic area, there may be a large number of distributors with responsibility for the individual stores. Thus the task of communicating a new chain level agreement to distributors can be complex. The supplier will want to ensure that every distributor is made aware of the agreement in a timely fashion. Equally important is to not clutter the information flow to a distributor when no individual stores are located within their distribution area.

When a distributor is serving multiple suppliers, the task of allocating their efforts to an individual supplier involves managing conflicting priorities. A distributor placed in successful situations by the actions of a supplier is more likely to give preference to that supplier.

The actions of the supplier and the involved distributors remain important over the entire term of the agreement. Additionally, there are factors that either limit the performance of the stores or influence the satisfaction of the chain and the stores that only appear well into the term of the agreement:

• The supplier has not reserved adequate quantities of the products involved so that sufficient inventory is available over the term of the agreement

• The distributors do not honor pricing over the term of the agreement

• The distributors do not maintain adequate quantities of the product available in local warehouses to prevent or limit the duration of out of stock incidents

These factors not only lead to lost sales but also may ultimately influence the decision of the chain to pursue additional agreements with the supplier. Often times the renewal of such an agreement is based on the level of service of the supplier and the involved distributors.

Communication is the Key

As illustrated in Table 1, there are a number of events along the supply chain that, if not managed, can result in lost sales. Table 1 describes these events in terms of success factors.

Table 1. Summary of Agreement Success Factors

Supplier	Distributors	Stores
Product is available for shipment	Product is ordered from supplier	Product is ordered from distributors
Product is shipped to distributors	Product is shipped to stores	Product is made available to consumers
Sufficient inventory is reserved for	Sufficient inventory is maintained	Sufficient inventory is maintained term of the agreement at local warehouses at each store
Agreed to pricing to distributors	Agreed to pricing to stores	Agreed to pricing to consumers is maintained is maintained is maintained

Communication among the various links of the supply chain is essential in order to maximize the sales opportunity presented by the agreement. Potential parties to the communication include the supplier, the headquarters for the chain, the individual stores, the distributors responsible for the individual stores, and the suppliers' local sales force responsible for the individual stores. A diagram showing the communication flow related to an agreement is shown in Appendix A.

The problem to be solved is how to easily provide a wide variety of information to individuals in a targeted fashion that ensures the right products are in the right place at the right time. The answer lies in building or buying systems that follow a data model where the relationships among all entities are correctly represented so that necessary communications and calculations are enabled.

An Explanation of the Data Model

The data model discussed in this section is depicted in Appendix B (on the next page) and should be thought of as a logical model even though it appears to include physical tables and columns. For example, several tables have a ContactInfo column. In an actual database contact information would be stored in several columns like address, city, state, zip code, e-mail address and so on. Additional tables may also be required in order to support multiple-valued information like phone numbers.

The ContactInfo column, and columns like it in the model, should be viewed as placeholders for the actual columns that would be necessary in a physical database. In a similar fashion, the tables in the model should be thought of as logical entities that may require multiple tables for implementation in a normalized database.

In database terminology, a primary key can be thought of as the column or columns that uniquely identify a row in a table. 'PK' in Appendix B notes a primary key. A foreign key, noted by 'FK', is a column in a table that is the primary key of a related table. For example, the primary key of ChainId and AgreementId uniquely identifies a row in the Agreement table. ChainId is also noted as a foreign key in the Agreement table since it is the primary key in the related Chain table.

To solve the problems identified in the previous sect is permissible for a store to feature a product but where each store may decide whether or not they participate. This will be called an optional feature and often occurs in franchised restaurant chains that leave certain operation decisions to the franchisee.

To record a mandatory feature it is not necessary to use the Store Product table. It can be assumed that all stores listed in the Store table would participate in the agreement. The exception would be if it were desired to record estimated quantities for a specific product at the store level.

A possible extension to data model would be to implement the Chain table as a hierarchy. One or more levels above chain could be used to capture parent-child relationships such as Darden Restaurants being the parent of both the Red Lobster and Olive Garden chains. Levels below chain could be used to support situations where a franchisee owns a number of stores or where a chain uses geography to subset its stores.

Implementing the Chain table as a hierarchy facilitates recording information about a subset feature. The agreement is not related to the chain but rather to a lower level in the chain hierarchy corresponding to the participating stores. Similar to the mandatory feature, it would not be necessary to use the Store Product table since it could be assumed that all stores in a particular lower level of the chain hierarchy would be participating. Absent such a chain hierarchy implementation, individual stores participating in a feature should be recorded in the Store Product table.

There is little flexibility in supporting an optional feature. Since each store may decide whether or not to participate, those stores choosing to participate must be recorded in the Store Product table.

Additional information that needs to be recorded in the Store table relates to the teams of representatives from both the supplier and the distributor that are assigned to each store. The Store Table includes foreign key references back to both the distributor team and the supplier team responsible for a particular store. Both the Distributor Team and Supplier Team tables are related to Rep tables that define the set of representatives on each team. This structure is designed to prevent having to update the information in the store level when the representatives on the teams responsible for the store are changed via termination or reassignment. In such a situation, only the associated Rep table needs to change.

Facilitate Communication Related to the Agreement

Over the term of the agreement, sales are maximized when the right products are in the right place at the right time. This maximization requires specific actions by the supplier, distributors, and the chain and its stores. As noted previously, efficient communication is the key factor in making sure that appropriate actions are taken.

The need for communication among the entities associated with an agreement can be triggered by a number of events. These events include the initial agreement with a chain, a featured product receiving acclaim from the press, a chain opening new stores or an inventory issue along the supply chain.

To maintain credibility with distributors and chains, a supplier must first avoid bad communication. Examples of bad communication include:

• Forwarding acclaim to a distributor not eligible to receive the product

• Sending details of an agreement to a distributor with no stores in their distribution area

• Not providing advance notification to a chain or its stores when inventory issues will impact the ongoing availability of a featured product

These examples can be avoided by using the relationships among entities in the data model to derive a set of contacts limited to only those individuals with responsibility for implementing an agreement.

To illustrate the process, suppose a new agreement with a chain is obtained and the relevant information has been recorded. To obtain the set of contacts related to the agreement, the processing logic is as follows:

1. Obtain the ChainId from the Agreement table

2. Use the ChainId to select contacts from the Chain Contact table

3. Use the ChainId to select a set of StoreIds from the Store table

4. For each StoreId, select contacts from the Store Contact table

5. For each StoreId, use the DistributorId and DistTeamId found in the Store table to select the responsible distributor representatives from the Distributor Rep table

6. For each StoreId, use the SupTeamId found in the Store table to select the supplier responsible representatives from the Supplier Rep Table

Once the set of contacts is obtained, it is a simple exercise to use the either the physical address or e-mail address of each contact to deliver the desired information.

Some communication may need to be limited, perhaps to only the impacted distributors or the headquarters of the chain. Narrowing the set of contacts can be easily achieved by modifying the processing logic presented above.

Information regarding inventory levels of featured products is another important communication topic. Presumably, the supplier would only enter into an agreement if an estimate of increased sales justified the costs associated with the agreement. Such an estimate can be that basis of calculations that help manage inventory.

In the data model, the QtyEstimate column appears in both the Featured Product and Store Product tables. The column in the Featured Product table is intended to capture an estimate of the total quantity of the product that will be sold due to the agreement. The column in the Store Product table is similar, except the estimate is by store. Yet a third option would be to extend the data model so that estimates could be captured by store and by time period so that seasonal variations are taken into consideration. The level of granularity to be used to capture estimates is an implementation decision made by the supplier.

Once estimates are recorded several calculations become obvious. Sufficient inventory of a product must be initially available to satisfy the quantity estimates across all agreements featuring the product. In the case of a product with limited availability, inventory should be reserved for distributors based on the number of stores in their distribution area that are participating in agreements featuring the product. The accuracy of these calculations depends on the accuracy and level of granularity of the estimates.

The data model includes a Shipment table intended to record shipments of a particular product to a distributor. This table enables a calculation that can uncover whether a distributor has received shipments sufficient to meet demand for stores in their area based on the estimates. Extending this calculation based on inventory or depletion data provided by the distributor can increase the accuracy of the calculation.

In the event that any of the calculations uncover an inventory issue relating to a product, the techniques described previously can be used to determine who should be contacted. Whether the calculations prevent a negative situation or facilitate communication after the fact, the credibility gained by demonstrating competence in managing agreements positions the supplier to be a supplier of choice.

Calculate Performance of the Agreement

In the absence of a calculation showing volume gains due to an agreement, determining the success of an agreement becomes a subjective exercise based on no true rate of return on the investment made in obtaining and servicing the agreement.

The Depletion table included in the data model is intended to be the historical repository for depletion data. A depletion is defined to be the transfer of a quantity of a product from a distributor to a retail account. Depletion data can be gathered directly from distributors or provided by a third party such as Beverage Data Network or Winery Exchange via their TradePulse^TM (more information is available at www.wineryexchange.com ) product.

The Sale table included in the data model is intended to be the historical repository for retail sales data. A sale is defined to be the purchase of a quantity of a product from a store. Sales data is provided by third parties such as ACNielsen (www.acnielsen.com ) or Information Resources Incorporated (www.infores.com).

Since sales data is sometimes based on estimates or sampling and since some chains choose not to divulge their sales data to third parties, this discussion will use depletion data as the basis of calculations. The assumption here is that a store would not receive product from a distributor if the near term sale of the product were unlikely. A supplier may choose to base their calculations on sales rather than depletions if the accuracy and coverage is sufficient.

The Product table in the data model shows Price and Cost columns. This is admittedly an oversimplification, especially in the area of price where prices can vary over time and by geographic area. A physical database would likely require several tables to support the complexity of this data.

With data described above and in previous sections, it is possible to perform a number of calculations related to an agreement. Because of the wide variety of possible physical implementations, the following are general descriptions of the process rather than exact processing logic:

• Get historical depletion data to determine pre-agreement volume level

• Get depletion data for periods associated with the term of the agreement

• Use year over year figures to account for seasonal variations

• Calculate change in volume

• Use product cost and price data to calculate gross profit attributable to the agreement

• Get cost data for events associated with the agreement

• Calculate net profit and return on investment

• Get agreement volume estimates

• Calculate variances between estimates and actual results

This discussion has shown that the data required to calculate and communicate information related to agreements may come from multiple systems, both internal and external. The presented data model includes columns, such as ChainId, StoreId, AgreementId, ProductId and DistributorId, that function as both primary keys and foreign keys across virtually every table in the model. It is essential that either a common set of column names and data values or robust translation mechanisms supported by a data dictionary exist in order to support the system functionality required to effectively manage agreements with chains.

Obtaining agreements with a chain is only the first step of a process. In order to maximize the sales opportunities presented by such an agreement, a complex communication flow must be managed by the supplier through all links of the supply chain. This will provide a level of service to the chain and its stores so that the chain is willing to enter into additional agreements with the supplier. Also, the projected sales of the agreement must be compared with actual store level depletion and sales data so that a true calculation of the profitability of the agreement can be made.

By creating a set of internal systems where data is integrated based on a consistent data model, a necessary service level can be maintained and the true profitability of an agreement can be determined. Such an integrated systems environment places the decision making process completely in the hands of the supplier so that profitable agreements can be maintained while non-profitable agreements can be terminated or made profitable. wbm