Patent Application
20140214581
Reference is made to commonly assigned U.S. patent application Ser. No. ______ filed concurrently herewith, entitled “Activation of Media Product Aggregation Using Order History” by Marcello Balduccini et al; U.S. patent application Ser. No. ______ filed concurrently herewith, entitled “Aggregation of Media Product Production and Distribution” by Marcello Balduccini et al; U.S. patent application Ser. No. ______ filed concurrently herewith, entitled “Aggregation of Customer Requirements” by Marcello Balduccini and U.S. patent application Ser. No. ______ filed concurrently herewith, entitled “Adjusting a Customer Catalog for Ordering Visual Media Products” by Kevin Gobeyn et al, the disclosures of which are incorporated herein in their entirety.
The present invention relates to methods for aggregating customer requests for the production and distribution of custom products and services to enable collaborative group buying.
U.S. Patent Application Publication No. 20120084143 A1 “Optimal List-Price Mechanism Design for Multi-Level Device Click-Through in Targeted Print or Electronic Communication”, is directed to delivering at least one communication from an advertiser to a consumer or in other words for direct advertising. It refers to an “advertisement product”, which is defined as “on-demand printed or electronic communications”, and it employs the notion of a “demand curve” to evaluate the demand for certain products. However, this demand curve considers the current demand and is not intended to anticipate demand fluctuations. It also does not address the phenomenon, common in production scenarios, of the reduction of the per-unit price when the total units for an order increase. The demand curve simply provides a “count” of how many advertisers are willing to pay a given price. It does not provide an account of the past or permit one to consider seasonal or regional variations in demand, which are often important when trying to anticipate such demand fluctuations. Moreover, it does not in general permit one to form reliable predictions about the future. It simply describes what the offers are right now. The ability to anticipate demand fluctuations is also a feature of the present invention and thus, to overcome the above limitations, the notion of “historical demand model” is introduced, based on the use of pricing grids and order history.
It should be noted that a further shortcoming of the prior art reference is that it does not have contingencies for producing and distributing “non-advertising” products or services such as photo books, greeting cards, custom wall paper, and personalized and short run calendars, apparel, and the like.
A feature of the present invention is that it gives equal importance to production costs and distribution costs. In contrast, most of the current art attempts to only minimize production costs, either by selecting only low cost production entities and processes by using a bidding process where the lowest bid wins, or by modifying product requirements, such as using lower cost materials, lower quality processes, format changes, or reducing sizes.
Some production entities rely on a technique referred to as “contribution pricing”, as described in U.S. Pat. No. 6,397,197B1 “Apparatus And Method For Obtaining Lowest Bid From Information Product Vendors”, where they provide low or no margin pricing in order to keep their operation running at near capacity as opposed to enduring periods where their production equipment and staff are under-used or inactive. This form of pricing can alienate regular customers that pay the standard price and can negatively affect pricing in general. In addition, small volume “fill in” work is often disruptive to normal production, requiring changes to production set ups and procedures which also impacts costs and productivity. There is no competitive “bidding” used in the present invention, since prices are determined from production and distribution pricing grids. Aggregated requests are sent to the selected production entities based on compatibility between the production and distribution capabilities, as expressed by the pricing grids, and aggregated requests.
The present invention employs a form of “Collaborative Group Buying” instead of the above described “Competitive Bidding”. Competitive bidding places the burden on the production entities and will always force them to find ways to reduce costs which can negatively impact quality, service, and profitability. In contrast, by aggregating multiple customer product or service requests and distributing them amongst a pre-established network of production and distribution entities, collaborative group buying enhances operational efficiencies, reduces the required numbers of setups, and makes possible longer production runs, resulting in less equipment and staff idle time and disruption. Because the production entities are operating closer to full capacity, the per-unit production costs are reduced which can improve quality, service, and profitability.
U.S. Pat. No. 6,330,542 “Automated Internet Quoting and Procurement System and Process for Commercial Printing”, provides a Delivery Zip Code which is used to select the nearest available print provider given the criteria entered by the print buyer and to compute the freight charges. This technique relies on production entities that are located near the distribution destination in hopes of reducing distribution costs. The production entity that is within the selected range from the distribution destination and with the lowest production costs is chosen. The problem with this technique is that distribution costs do not necessarily increase monotonically or the productions entities closest to the distribution destination might not have the lowest distribution or production costs. For example, a production entity that is close to a private or public sector distribution hub as opposed to regional proximity to the destination would have greater access to lower cost distribution. In addition, a potential production entity might be located in a so called “Right to Work” state where lower cost and higher productivity non-union labor is readily available.
From a technical perspective, in order to simultaneously consider all of the factors involved in determining the appropriate production entity capable of producing and distributing a given aggregated set of customer requirements, procedural programming can be used to implement a reasoning system that takes all the production and distribution factors into consideration. However, this approach would involve a significant effort requiring tens of thousands of lines code, and the corresponding algorithms could have to be modified quite substantially whenever a new constraint or feature of the system is required.
On the other hand, an aggregation system implemented using declarative programming is more adaptable and can be more efficient as real world data and use cases are collected. At the same time, this type of approach still provides the potential to migrate from declarative programming techniques to ad-hoc algorithms implemented using procedural programming once the problem space is well understood, in order to achieve a further level of improvement.
U.S. Pat. No. 7,788,143B2 “System and Method for Competitive Pricing and Procurement of Customized Goods and Services”, describes a method for selecting a lowest bidding vendor from a plurality of vendors of a customized good or service. Vendor attributes from each of the plurality of vendors representing their respective capabilities but no distribution or pricing grids are provided. The vendor attributes or the invitation-for-bid, or both, are received through a web browser. The invitation-for-bid is compared to each of the vendor's attributes according to certain standard or optional selection criteria to produce a vendor selection pool of vendors qualified to bid on the job.
Each vendor in the vendor selection pool receives a vendor's invitation-for-bid. A bid is received from at least one vendor in the vendor selection pool, the lowest price bid is identified, the buyer is informed of the identity of the selected vendor, and solicited for approval of the selected vendor. Upon receipt of approval from the buyer, an order is issued to the selected vendor. The non-selected vendors in the selection pool are informed of the bid prices and of the selection results. The prior art system does not aggregate customer orders to produce higher volumes of similar products and services. This places an additional burden on the production entities, requiring shorter production runs, more set up time, and under-used production equipment, facilities, and staff. Also the prior art system does not simultaneously process customer requirements with at least production and distribution pricing grids.
A method is disclosed for managing production capacity comprising,
using a communication network to receive data including production and distribution pricing grids from a network of production entities;
receiving quoting requirement over the communication network from a plurality of customers for producing and distributing customer designated media products;
calculating a baseline of aggregated media product volumes and an associated target cost that are compatible with the capabilities and capacities of the network of production entities;
using a processor to determine if the current volume of received quoting requirements will produce the baseline of aggregated media product volumes;
calculating reduced target cost estimates when received quoting requirements are below the baseline of aggregated media product volumes;
calculating increased target cost estimates when received quoting requirements are above the baseline of aggregated media product volumes;
using the communication network to send the quoting requirements and the resulting calculated target cost estimates for producing the aggregated media products to a subset of the network of production entities;
receiving actual cost estimates based on the calculated target cost estimates over the communication network for producing and distributing the aggregated media products from the subset of the network of production entities;
calculating customer prices for the plurality of customers for producing and distributing customer designated media products based on the received actual cost estimates; and
using the communication network to provide customer prices to plurality of customers for producing and distributing customer designated media products.
These and other aspects, features, and advantages of the present invention will be more clearly understood and appreciated from a review of the following detailed descriptions of the preferred embodiments and appended claims, and by reference to the accompanying drawings.
The aggregator isolates the customer from the production entities and acts on behalf the customer reducing the level of involvement required by the customer. The pre-established network of production entities only needs to establish and maintain communications with the aggregator.
There are fewer job orders and billing requirements for the same volume of work for the production entities since they only have to interact with the aggregator and larger aggregated orders. Moreover, if the production entities are paid by the aggregator, then the billing and collection requirements of the production entities are reduced significantly.
By aggregating multiple customer product/service requests and distributing them amongst a pre-established network of production and distribution entities, operational efficiencies are enhanced, the required numbers of setups are reduced, and longer production runs are possible, resulting in less equipment and staff idle time and disruption. Because the production entities are operating closer to full capacity, the per-unit production costs are reduced.
The printing and publishing market is affected by the recent innovations in low-cost and service-provider price-subsidized handheld, portable devices with real-time wireless access to content via the internet and cellular networks. As a result long printing runs required for magazines, catalogs, newspapers, and the like are in decline. By providing a system that collects and aggregates multiple customer orders for similar products and services and distributes these aggregated orders to a pre-established network of production entities the operational efficiencies, quality, and service provided by the production entities are maintained in a consolidating market.
Members of the pre-established network of production entities are certified as part of an “on-boarding” process. The aggregator is unconcerned with the detailed capabilities and day-to-day operations of the production and distribution entities such as the types of equipment, maintenance schedules, ongoing training, staffing levels, facilities, and so on. The aggregator is concerned with the supported product types and formats, volume capacities, and services supported by the production entities and communicated via the production and distribution pricing grids production and distribution pricing grids and response times for requests for estimates where appropriate.
There is no competitive “bidding” used in the present invention. Prices are determined from production and distribution pricing grids. Aggregated requests are sent to the selected production entities based on compatibility between the production and distribution capabilities, as expressed by the pricing grids, and aggregated requests. The present invention employs a form of “Collaborative Group Buying” as opposed to the more common “Competitive Bidding”, which instead places the burden on the production entities and will force them to find ways to reduce costs. This often negatively impacts the range of products supported by the production entities. It also can decrease the quality of the products, the service provided or the profitability of the production entities.
The invention is more completely understood by considering the detailed description of various embodiments of the invention which follows in connection with the accompanying drawings. Referring now to the drawings in which like reference numbers represent corresponding parts throughout:
The aggregator stores historic information related to the timing, types, and volumes of prior completed customer quotes and uses the stored historic information to determine if the quoting requirements should be aggregated now or at some point in the future. This information can also be used to predict seasonal and regional variations in demand. For example, a resort or vacation destination will have increased demand for marketing communications, promotional materials, pamphlets, mementos, and souvenirs prior to and during the peak season. Purchasable memorabilia such as calendars, posters, post cards, totes, hats, t-shirts, and the like are often customized to incorporate colors, graphics, images, and text expressing the theme and points of interest of the local venues as is true for marketing communications for accommodations, local entertainment offerings, and recreational venues. The amounts, types, and distribution patterns of these marketing communications, promotional materials, mementos, and souvenirs are extrapolated from past orders and sales. Special holidays, social events, regional events, anniversaries, and celebrations, political campaigns, or any other recurring or planned event are used along with past production and distribution patterns to predict and anticipate demand.
In addition, demand forecasts can include simple “averages” of orders received on a daily basis, which are used to accumulate a history demand record and to identify patterns of demand. For example, a given product, such as custom postcard, typically is requested on a routine basis with unit volumes averaging 50 per day.
Other important aspects of the aggregator are the quoting and estimation deadlines, which will have to be considered in order to meet production and distribution deadlines. The quoting requirements are accumulated as long as there is enough time remaining to meet quoting, estimating, production, and distribution deadlines, and these deadlines are determined from the quoting requirements, the historical data, the production and distribution pricing grids, or combinations. In addition to historic demand data, data on the distribution locations can also be accumulated and used to predict a range of distribution costs that are anticipated and used to make more accurate and cost effective aggregations. This is accomplished using United States Postal Service (U.S.P.S.) ZIP codes or Postal Zone Charts, or any other standardized distribution table provided by a public or private distribution entity.
Production and distribution are provided by the same entity or by separate entities working collaboratively. Such would be the case of a commercial printer that uses U.S.P.S. for final distribution. The commercial printer provides presorted mail in approved containers based on the ZIP code destinations on the mail, in exchange for lower postage rates.
The U.S.P.S. also provides Business Mail Entry Units (BMEUs) which are areas of postal facilities where production entities present bulk and permit mail for acceptance. The BMEUs include dedicated platform space, office space, and a staging area on the postal workroom floor for use by production entities. The U.S.P.S. also provides Advance Deposit Accounts or “Trust Accounts” where a debit account is provided into which a production entity deposits funds that are maintained by the U.S.P.S. and from which postage is later deducted at the time of mailing.
As the described arrangements illustrate, the production entity can collaborate very closely with a distribution entity. In addition, with products that have standard formats and properties, such as post cards, business cards, calendars, photo books, photographic prints, catalogs, pamphlets, and CDs and DVDs, the shipping weights of these products are calculated directly from the product volumes using known unit product weights. The following U.S. patent Nos. describe these techniques and are incorporated by reference: U.S. Pat. No. 6,943,867 B2 Method and System For Shipping Of Photofinishing Orders, U.S. Pat. No. 6,829,037 B2 Method and System For Verifying A Photofinishing Order, and U.S. Pat. No. 6,812,998B2 Method and System For Shipping Of Photofinishing Orders.
A production entity also referred to as a Production Service Provider or “PSP”, might be able to fulfill an order for 500 units by the production or distribution deadline but not an order of 5000 units. In addition, PSPs can specify multiple production and distribution time frames to accommodate special circumstances such as a premium priced “Rush Order”. These special circumstances can also be included in the pricing grids.
The following scenario is an illustrative example of a typical routine Annual Catalog Order from a regional retailer. For the previous 6 years “Acme Fine Apparel (AFA)” has had their annual Mother's Day Sale. 30 days prior to the sales event AFA has requested quotes for producing catalogs and distributing them via a bulk mailing to several target zip codes within a 10 mile radius of their 3 retail locations. In the first year AFA had one retail location and had requested quotes for 10,000 catalogs, but over the six year period the business had expanded to three retail locations and last year ordered 40,000 catalogs, each containing 12 full color pages with an inserted page of coupons. The prior years' orders were analyzed and according to the trend an order of between 40,000 and 45,000 catalogs, distributed to 3 zip code locations via bulk mail, and each containing 12 full color pages with an inserted page of heavy stock, pre-perforated coupons. This historic demand data is used by the aggregator to anticipate these order types, volumes, and distribution patterns and a time frame to expect to receive a request for quote from the potential customer.
Another example involves Seasonal Greeting Card Orders from individual customers. Each year for the past five years, at 6 weeks prior to the Christmas holiday, orders for holiday cards start to arrive. Each week the number of orders increases, peaking one week before Christmas and dropping to pre-holiday levels a week after Christmas. Individual orders range from between 25 cards (minimum order) to approximately 500 cards, with each order requesting selected digital art work.
Historical demand provides the basis of demand forecasting. Obviously, no demand forecasting technique is completely accurate so multiple approaches can be combined to augment or replace simple historic averages. Combining forecasting methods improves accuracy and reduces the chances of large errors.
There are many well-known demand forecasting techniques that are used within the present invention such as “Rule-Based Forecasting”, where normalized data is processed using short and long range models. For each of these two models, trends are identified; for the long range model, trends are dampened due to the greater uncertainty caused by the longer the time horizon. Results from both the short and long range models are then combined to obtain a rule-based demand forecast.
Another approach to demand forecasting is the use of mathematical models such as “Neural Networks” for modeling the complex relationships between historic information such as order quantity type and volume, previous sales, time of year, order locations, or delivery locations. Neural networks are advantageous in that they are adaptive systems that can change their structure during the learning phase of deployment.
“Data Mining” is another technique used for demand forecasting and involves semi-automatic or automatic analysis of large volumes of data to extract previously unknown patterns. Detected patterns include: groups of data records referred to as a cluster analysis, unusual records or anomaly detection, and dependencies know as association rule mining. These patterns are summaries of the accumulated historic data and are used to identify multiple groups in the historic data, which are then used to obtain more accurate demand forecasting.
“Causal Models” rely on the assumption that it is possible to identify the underlying factors that influence the value of a variable of interest. Forecasts based on linear relationships between identified variables that occur for long periods of time are useful in predicting such a relationship in the future. In addition, non-linear repeating patterns such as seasonality as previously described in the “Annual Catalog Orders” and “Seasonal Greeting Card Orders” examples can also be used with this technique. The “Time Series Projection” methods also use historical data as the basis of estimating future outcomes. These methods include “Simple Moving Average”, popular for instance in predicting stock market fluctuations, which relies on the arithmetically averaging of past data over a specified period to project forward in time in order to smooth out short term fluctuations and highlight long term trends or cycles. This is a very simple way to implement smoothing algorithm but is generally considered to provide low accuracy results.
Another Time Series Projection method is “Exponential Smoothing”, which is also a smoothing technique that is applied to time series data. This technique is used to smooth data or to make forecasts. Compared to the Simple Moving Average technique where past observations are weighted equally, exponential smoothing assigns exponentially decreasing weights over time.
Yet another Time Series Projection method is the “Trend Projection” method, which is focused on observing the magnitude and direction of the movement of a variable through time. This method requires reliable long term time series data.
The aggregator receives quoting requirements from customers and provides quotes to customers. In this sense, the aggregator acts as a liaison between are between customers and the production service providers (PSP). The aggregator also provides requests for estimates, derived from quoting requirements provided by the customers, to the production service providers and in response receives estimates from the PSPs. Quoting requirements and quotes reflect the price to be paid by the customer for products and services requested and requests for estimates and estimates reflect the price to be paid to the PSPs for the products and services provided.
The first embodiment of the present invention provides a method for the activation of media product aggregation based on order history.
The above information is used by the selection component 170 in order to determine which QRs 120 are immediately forwarded to the aggregation component 190 and which ones are held for aggregation in the future.
Selection component 170 behaves as follows. First, based on historic information 40, the aggregator 10 determines the expected total volume, volume(p,m), of product p requested in QRs 120 received on a given day, and the expected average quoting deadline of such QRs 120. As previously described various demand forecasting techniques or combinations of those techniques are used.
Let daily_orders(p) denote the expected total volume of product p requested in QRs 120 received per day and let avg_deadline(p) denote the expected average deadline of such QRs 120. These values are obtained from the historic information database 30. Those skilled in the art will see that it is not difficult to extend the formulas below in order to consider different values of daily_orders(p) and avg_deadline(p) for each day. Intuitively, each of the QRs 120 that are expected for day i will become due avg_deadline(p) days after dayi, and will have to be assigned. Therefore, the corresponding volume of product will no longer be on order.
Hence, the expected total volume of product p on order on day m (m≧0, with 0 meaning today) is:
where volume(p,0) is the quantity of product currently on order, due (p,i) is the total volume of product p in QRs 120 that have already been received by the system and due on day i, and
A QR for product p is at peak for production on day D if on every day between today and the day the quote for QR is due, the expected volume on order for p is less than or equal to the volume on day D.
Next, historic information 40 and the received QRs 120 from quoting requirement database 90 are used by the aggregator 10 to determine the expected total shipment weight (weight(m,l), for a day m and a final destination l).
Let daily_orders_for_destination(p,l) denote the expected volume of product p in expected QRs 120 with final destination l and avg_deadline(p,l) be their expected quoting deadline. These values are obtained from the historic information database 30. Once again, those skilled in the art will promptly see how such estimates are made to depend on a given day.
The expected total shipment weight of product p for day m with final destination l, denoted by weight(p,m,l), is:
where weight(p,0,l) is the weight of product currently on order, due_weight_f or_dest(p,i,l) is the total weight of product p in QRs with destination l that have already been received by the system and due on day i, and
The expected total shipment weight for day m and final destination l is then:
The QR 120 is at peak for mailing on day D if on every day between today and the day the quote for QR 120 is due, the expected shipment weight of the QRs 120 for the corresponding destination is less than or equal to the expected shipment weight on day D.
On any day D, the QR 120 can be at peak for production, for mailing, or for both. The system permits the operator to specify a preference order among the peaks. The preference order specifies which peak is better than another. The QR 120 is at an overall peak on day D if it is not expected to reach a better peak on any other day.
The QR 120 is at the latest overall peak on day D if it is at an overall peak on day D and it is not expected to be at an overall peak after day D.
Finally, the QR 120 is selected for aggregation on the current day by selection component 170 if it is currently at the latest overall peak. The set, 180 of QRs 120 that are selected for aggregation is then forwarded to aggregation component 190.
Those skilled in the art will see that selecting for aggregation the QR 120 when it is at the latest overall peak is appropriate when the frequency of withdrawal of QRs 120 by the customers 80 before their quoting deadline is expected to be low. Those skilled in the art will see how the selection criterion is modified to accommodate for more frequent withdrawals of QRs 120.
The second embodiment of the present invention provides a method of evaluation of media product aggregations based on order history.
A more detailed explanation is provided by
Finally, the aggregator 210 also retrieves from the historic information database 30, information regarding the timing of prior completed customer quotes. Every time QRs 120 are received by the aggregator 210, the corresponding information is used to update the historic information database 30.
The above information is used by the selection component 170 in order to determine which QRs 120, should be processed immediately by the aggregator 210, and which ones should instead be held for aggregation in the future.
In this embodiment of the present invention, an aggregation component 2270 computes aggregations of the QRs 120 received from the customers 80. The computation is performed under the assumption that the received QRs 120 will be processed immediately by the aggregator 210 (as opposed to being held for aggregation in the future). The aggregation task is performed as illustrated in
On the other hand, suppose that historic information database 30 indicates that QRs 120 for a total volume of 750 pages of p1 are expected to be received by the end of the day. The corresponding anticipated QRs 250 can thus be fed to aggregation component 1260 together with the information about the current QRs 120. It is not difficult to see that aggregation component 1260 can now find an anticipated aggregation 230 for p1 with a total volume of 1,000+750=1,750 pages. This aggregation, if the anticipated QRs 250 are indeed received, gives access to the discounted unit price of $0.85 per page. In this example, then, it is advantageous to hold the current QRs 120 for p1 and wait until the anticipated QRs 250 are indeed received—or until a determination is made that the current QRs 120 can no longer be held, for instance because their quoting deadlines are about to expire. The task of comparing current aggregations 220 and anticipated aggregations 230 and of making the determination whether any current QRs 120 should be held is made by selection component 170. To make this determination, selection component 170 calculates one or more scores for each of the current aggregations 220 and anticipated aggregations 230. The scores measure various aspects of the quality of the aggregations, including price, expected job quality, features of the corresponding production/distribution entities 20, features of the corresponding shipping methods, and timeliness (i.e. ability to match the deadlines). The scores are then compared using methods known in the art and the aggregations with advantageous scores are selected. When the selected aggregations belong to the set of anticipated aggregations 230, the corresponding current QRs 120 are put on hold. When the selected aggregations belong to the set of current aggregations 220, quotes 130 are produced for the corresponding current QRs 120 using the information from the selected aggregations. Finally, the quotes 130 are sent to the customers 80.
The third embodiment of the present invention provides a method of determining pricing modifications in media product aggregation systems.
The above information is used by a baseline estimator 330 to calculate a baseline of aggregated media product volumes 350 that are compatible with the capabilities and, optionally, with the capacities of the network of production/distribution entities 20. The baseline 350 provides target volumes of QRs 120 that should preferably be received from customers 80. For example, the baseline 350 can indicate the volume of QRs 120 that is needed to reach the largest cost savings based on production/distribution pricing grids 50 and on a system capacity 400 produced by a system capacity calculator 390. The system also retrieves from the historic information database 30 historic information 40 regarding the timing of prior completed customer quotes 110; the information is used to determine a baseline 350 indicating the volumes of QRs 120 that are typical for the current period of the year. In this embodiment, every time QRs 120 are received by the system, the corresponding information is also used to update the historic information database 30.
Concurrently, aggregation component 190 receives the current QRs 120 and the production/distribution pricing grids 50 from the production/distribution entities 20. The information is used to compute aggregations of the QRs 120 and to calculate corresponding initial quotes 130 for the QRs 120. In a preferred embodiment of the present invention, the aggregations are computed according to the method illustrated in
The quotes 130 are then received by a target adjustment module 340 together with the baseline 350 and the QRs 120. The quotes 130 include an indication of the corresponding unit price for each of the QRs 120, which represents a target cost estimate 360. Target adjustment module 340 compares the volumes of received QRs 120 with the baseline 350 of aggregated media product volumes 350. If the amount of QRs 120 is below the baseline 350, then the target cost estimates 360 from quotes 130 are reduced. Intuitively this is done in order to increase the chances that the customers 80 will accept the quotes 130. If the amount of QRs 120 is above the baseline 350, then the target cost estimates 360 from quotes 130 are increased, which prevents overwhelming the network of production/distribution entities 20. Corresponding revised target cost estimates 370 are then sent to the price verification module 290. In a preferred embodiment of the present invention, the computation performed by target adjustment module 340 also takes into account the current system capacity 400, as calculated by system capacity calculator 390 according to the method discussed in the description of
Next, revised target cost estimates 370 are sent to the price verification module 290. The price verification module 290 confirms the revised target cost estimates 370 with the production/distribution entities 20, by sending to the production/distribution entities 20 suitable RFEs 150 and receiving corresponding estimates 160, as discussed in more detail in the description of
In order to ensure fairness towards the production/distribution entities 20, in a preferred embodiment of the present invention, the price verification module 290 selects the set of production/distribution entities 20 that should receive RFEs 150 according to information about the timing about past RFEs 150 and depending on how the target adjustment module 340 has revised the target cost estimates 360. If the revised target cost estimates 370 are lower than the target cost estimates 360 from the quotes 130 produced by Aggregation component 190, then production/distribution entities 20 are selected, which have not received RFEs 150 with reduced cost estimates for a predetermined period of time. Alternatively, the production/distribution entities 20 are selected in sequential order. On the other hand, if the revised target cost estimates 370 are higher than the target cost estimates 360 from the quotes 130 produced by aggregation component 190, then production/distribution entities 20 are selected, which have not received RFEs 150 with increased cost estimates for a predetermined period of time. As above, the production/distribution entities 20 can also be selected in sequential order. This method ensures that production/distribution entities 20 with suitable capabilities are queried in a fair way, preventing the aggregator system 310 from sending RFEs 150 involving increased cost estimates always to the same production/distribution entities 20, which would receive an unfair advantage. Similarly, this method also prevents the system from sending RFEs 150 involving reduced cost estimates always to the same production/distribution entities 20.
The fourth embodiment of the present invention provides a method for efficiency enhancement of product aggregation using declarative programming techniques.
In addition, the production/distribution pricing grids 50 provided by the pre-established network of production/distribution entities 20 includes additional information about the production/distribution entities 20 that can be important to potential customers. This ancillary information about the individual production/distribution entities 20 includes various business and industry designations such as minority owned status, women owned status, exclusive customer relationships, environmental compliance status, unionized workforce, non-unionized workforce, ISO certification, and energy efficiency designation. Customers 80 can request, via their QRs 120, that one or more of these selection criteria be used in selecting an appropriate production/distribution entity 20. The production/distribution pricing grids 50 provided by the pre-established network of production/distribution entities 20 also includes available substrate types and substrate width capacities. This information is useful for determining production capacities and for determining the types and the range of custom products that are produced.
In
The above information is used by the aggregation component 190 in order to determine (1) which QRs 120 are assigned to a common production/distribution entity 20 in the form of a single order, rather than separate orders, in order to increase production and shipping volumes and receive better per-unit prices from the production/distribution entity 20; (2) which production/distribution entities 20 should receive the aggregated orders. These two steps are listed separately in this description of the aggregation component 190. Nonetheless, the two steps above can occur sequentially or in parallel, without affecting the present invention.
Let us introduce some terminology. Let Q be the set of QRs 120 that have to be processed by the aggregation component 190. An aggregation is a set A⊂Q such that the QRs 120 in A are assigned to a single production/distribution entity 20. An aggregation scenario (scenario for short) is a set S of aggregations from Q such that (1) the aggregations in S are pair wise disjoint; (2) every QR 120 from Q occurs in some aggregation in S; and (3) no two aggregations in S are assigned to the same production/distribution entity 20. The task of the aggregation component 190 includes (1) determining a set of scenarios and (2) identifying among those the most favorable one or ones, according to pricing, customer, system and operator selection criteria, and other system parameters as appropriate. When the system involves a human operator who is in charge of making the final decision it is appropriate to return multiple aggregation scenarios. The operator can then consider factors that are unknown to the aggregation component 190 when making the final decision. When the system acts without the intervention of a manual operator, it is more appropriate to identify a single most favorable scenario. The above two tasks can occur sequentially or in parallel, without affecting the present invention. Those skilled in the art will promptly notice that many alternative aggregation scenarios for a given set of QRs 120 are possible.
The second case from
Optionally, scenario selection module 490 can send select scenarios as scenarios to be verified 520 to the price verification module 290 in order to determine accurate prices for the aggregations contained in those scenarios. The updated scenarios returned by the price verification module 290, if any, are then used by the scenario selection module 490 for the ranking of the scenarios. The price verification module 290 will typically perform its task by sending suitable requests for estimates (RFEs) 150 to the production/distribution entities 20. This permits the production/distribution entities 20 to adjust the prices in response to sudden, temporary changes in the production environment, such as unexpectedly low demand. The production/distribution entities 20 are expected to respond to the RFEs 150 with estimates 160, which specify updated prices or alternative offers. It should be noted that giving the production/distribution entities 20 the ability to adjust their prices at this stage does not compromise the system's fairness. Of particular concern is preventing production/distribution entities 20 from entering unrealistic low prices in their production/distribution pricing grids 50 in order to be selected by the system, and then increasing the prices when the price verification module 290 requests estimates 160 from them. This situation cannot arise in the system described here, because the revised scenarios are re-ranked by scenario selection module 490 after they are received from the price verification module 290.
Next, the best scenario selection module 500 selects a best scenario 505 from the most favorable scenarios 530. In the simplest case, the selection is performed manually by an operator, as shown in
Finally, the best scenario 505 is used by quote generation module 510 to prepare and send quotes to the customers 80. The challenge of the task performed by quote generation module 510 is in spreading in a fair manner the savings obtained by the aggregations in the best scenario 505. A simple approach in determining a per-unit price from each aggregation of the scenario is to obtain a price for each QR 120 in the aggregation by multiplying the per-unit price by the volumes specified in that QR 120. Margins and fees can also be added, for example as a percentage or as fixed values. In this approach, the savings obtained by aggregating the QRs 120 are spread equally throughout the aggregated QRs 120. A possible alternative approach would be to assign to each QR 120 an amount of savings that is proportional to the volumes specified in that QR 120. This permits customers to be awarded cost savings that are proportional to how much their QRs 120 have “contributed” to achieving the overall cost savings. Customers 80 with the largest-volume QRs 120 in an aggregation would receive the largest savings; customers with smaller-volume QRs 120 would receive smaller savings, but they would still save money compared to the rates they would have received on their own. The quotes are then sent to the customers 80.
It should be noted that, in order to reduce the amount of scenarios considered by aggregation component 410, and thus improve performance, it is possible and often advisable to execute some of the modules in
The fifth embodiment of the present invention provides a method for using the aggregator to dynamically update customer facing catalog pricing.
The above information is used by aggregation component 190 to compute aggregations of QRs 120 and calculate a corresponding initial pricing for QRs 120. In a preferred embodiment of the present invention, the aggregations are computed according to the method illustrated in
where capacity (e,p) denotes the capacity of production entity e for product p, measured in terms of the volume of orders for p that e can still accept. In another preferred embodiment of the present invention, system capacity 400 for product p is calculated by taking into account the influence of existing orders for products different from p. Consider for example the case of a production entity e1 that has a single press capable of producing product p and some other product p′. If the press is already in use for the production of a large order for product then capacity (e1, p) is lowered accordingly. Such adjustment can be performed in several ways. In one such method, fixed coefficients are provided to the system. A coefficient α(p′,p) indicates how much the current production of p′ affects the capacity of p. The initial value of capacity(e1, p) can then be lowered by α(p′,p)·volume(e1,p′), where volume(e1, p′) is the volume of product p′ already on order. Another method for adjusting the capacity takes into account the deadlines of the orders. For example, the production of an order that is scheduled to be completed in the next one or two days is unlikely to overlap with the production of an order that is currently quoted. An adjusted value of capacity (e1,p) can thus be obtained by taking into account the existing orders that are due at least k days in the future, where k is a system parameter that acts as a threshold. In yet another embodiment of the present invention, system parameters including α(p′,p) and k are computed using machine learning and data-mining techniques, known in the art, such as neural networks.
Given the initial pricing 650 and the system capacity 400, a price adjustment module 660 adjusts the unit pricing in response to the system capacity 400 and sends correspondingly adjusted quotes 670 to customers 80. In a preferred embodiment of the present invention, the price adjustment module 660 reduces the prices when system capacity 400 is high, in order to attract orders, and increases the prices when system capacity 400 is low. Capacity thresholds clow and chigh specify which beyond which values system capacity 400 should be considered, respectively, low or high. Coefficients adjlow and adjhigh specify by which percentage prices should be decreased or increased depending on the system capacity 400. In another embodiment of the present invention, the amount of the price adjustment is proportional to the difference between the system capacity 400 and the corresponding capacity threshold. In yet another embodiment of the present invention, the initial pricing 650 obtained from aggregation component 190 includes a margin charged for the aggregation task. The price adjustment performed by price adjustment module 660 modifies the margin component of the initial pricing 650, as described earlier. In another embodiment of the present invention, the margin and non-margin components of the initial pricing 650 are adjusted in different proportions, according to coefficient provided to the system.
In addition special data 770 as previously described, representing soft and hard constraints such as minority owned status, women owned status, or exclusive customer relationships supported by each individual production entity, are presented in the final column. The code key representing special data 770 is provided by data key 780 at the bottom of the UI 790. UI 790 provides the operator, not shown, of the aggregator system with a simple, easy to read representation of the production/distribution pricing grids 50 provided by the production/distribution entities 20.
Aggregation 11120 includes QR IDs 700 r1, r2, r3, r10, and r6 is shown with similar destinations 740 as expressed by similar U.S.P.S Zip Codes. The selected production identity ID 800 for aggregation 11120 is “Universal” and the selected QRs were aggregated primarily by delivery destination 740. In contrast, aggregation 21130 has the same product type data 710 P1 requested by QR IDs r4 and r5 and PSP ID 800 indicates that “Acme” was selected to fulfill the indicated QRs 700. Also note that the destinations 740 aggregation 21130 are different indicating that aggregation 21130 was aggregated primarily by product type data 710. No production entity has been selected for aggregation 31140 as illustrated by the “HOLD” message in the production identity ID column 800 indicating that more compatible QRs are anticipated for this aggregation.
Any or all of the described User Interfaces can be used with the described embodiments and can be used to monitor the aggregator system or to interact with it.
The present invention has been described in detail with particular reference to certain preferred embodiments thereof, but it will be understood that variations and modifications can be effected within the scope of the invention.
