In addition to actual transactions that came from several information systems of the four companies, each supply chain member provided data about lead Times and costs of the activities to be included in the optimization model, the franchisor also provided the forecast used for the LTO. This section includes a description of the forecasting process used by the franchisor, ihe forecast data collected, ong the description of transactional data that came from information sysiems.
rorecast
The forecasting process that the franchisor used for LTOs had two paris. There was a quantitative forecast process that was used to predict total weekly sales for each restaurant. The other part was provided by the marketing managers at the franchisor. They produced an “expert's forecast” which was the expected
134
MO|4 JONPOIg Sus Jo dow ulpyD Ajddng 2IuIDuÁq D BuIID912 Tm>; oinBI4
S19U101Sn2 pu3 12U1O SuE1n81S9H 19U1O
Wx (9/8611) ›:o— [ẹE Liền 19
c‹ sisjeue 84} Ul Papnjoul 0u s@pON
(09) 2189 ar at 90ÿ9'LỆ ơ--< 1euoisn2pua (yi) ổ- 2q jeuoibay
+— (ZLLZ) ZEL (€yl) 99L ` (LL) €68¢ — ()01Ƒ—— op | —— (§8E2) e202 899 6Z 0z9 (1) 08 8# 2q 1@U101Sh2 pua Kee) | () 0E ~4— (SỳSy) 68Z (0£) 806————— (2) 96yy— ) F |088£) 099E
a L£69'€ a beg J19u10†sn2 pua OL Tứ L —.... Ue) 6z _ 1) 0921 — —— (BI FE—— 691) ZZ2I 19uI0Jsn2 pua Kia ; ¿L€y Yš 4 9q
<— (0ss1) 28 (26) S6 ee (11) 0891 — —8)£ [_—— W881)98. (Lp) 98y/ ZL
, t9 (y2) 618EL Z - 9q FuofSu JuIosn2 pua Vi L 69y/ 05 0Z - 9# © Lf rw <— (sev) 061 G12) 91z-—— = (1) ogee — K 62) J F— (SE) I/8E (es) leszz G8/Z}tL oT 0 19uIosn2 pug Sứ ẹ ¿ -y# 3q +(81) (€8) §8— (8) 8/Iz——] — (re F—— (6002) 802 19uI0Jsn2 pua Kg) ẹ ¿Ê66 9 F—tĐI 57 -Ê#oq (L) ev ~<— (9922) 781 (981) 6L —— tes) €6) 9182 —)y X01) 0211 M (90) 91L Y` . l08VlL Sr- vouiorsno Pua tay / -ứ 9q () 80 ae - 2q |euolbe ~4— (E99) 9y (i08) weq————è— (2) 8 - 9q [euolbey
y2)8 [7 (9166) 80h —— M (8L) zLes Jawoysnd pug Xu ; L986'8/ £7- I#2q
dc mm 2289) 282 (ZL) ple res (St) 1696 —— (1/3/2743
+ (SueuidIus Jo 1equInN) p9ddIus sese2 J9AO 1/8] S9SE2 G[Š,8B5N - 9uIEN S,900N
Key
0 OF - spoos paysiuly
(62) 8Ly98
TL - S|EH91EIN EM
(2) 8Ly96 — Oo
135
percentage of total sales produced from promotional meals. Since the expert's forecast is expressed as a percentage of total sales, if was referred to as LTO Mix.
This section includes a description of the two parts of the forecast for the LTO, as well as the planning process.
Forecasting Total Sales for Restaurants. [he forecasting orocess for total sales was based on the moving average method. Each restaurant manager produced a forecast as follows:
® Add weekly total sales for the last 6 weeks.
@ Subdtract ihe highest and the lowest weekly jotal sales numbers.
@ Divide the resulting number by 4.
Using a moving average does not seem to be the best fit with this business environment. Restaurants’ total sales showed trend and seasonality, which is an indication that other techniques, such as exponential smoothing, might perform better. However, a moving average forecast is economical and simple which enables restaurant managers fo produce their own forecast independenily.
Expert's Forecast (LTO Mix}. Based on marketing intelligence such as past expernence, focused groups, and test markets, marketing managers at the franchisor developed the LTO mix. The LTO mix was the expected sales of the promotional meals as a percentage of tofalsales. The LTO mix was a single percentage number for ail restaurants for each phase of the LTO, from initiation to termination, Each ohase lasied a week, and from initiation fo termination there are eight phases in a LTO.
Itisimportant jo note that a LTO does not have a “IIH” effect on restaurants’
total sales. This was expressed by the franchisor’s management and corroborated by the data. LTOs were a key component of the success of the business and, virtually, there was a promotion running all the time. If the promotion did not include a
136
specifically developed meal, it was called a “value meal”. The resul was not an actual “Hit”, but a fraction of the fotal sales of a restaurant was in the form of the LTO du jour.
Planning Limited-time Offers across the Supply Chain. Affer producing the LTO Mix, a marketing manager from the franchisor issued a memo to ail restaurant managers and suppliers involved in the LTO. in this memo, the details of the LTO were specified, The memo included the characteristics of the LTO such as name, suppliers involved, ingredients and portioning, cost of ingredients and selling price.
The elemenis of these memos of interest to this section were the forecast LTO Mix by phase (or week} and the forecasi of total sales for an average restaurant. The forecast of total sales for the average restaurant was used only for the initial pipeline fll, Affer this formal communication, ail coordination among the supply chain members was informal.
The memo was sent fo all affected managers about 10 weeks before the LTO was initiated, and if is used by manufacturers and distributors to plan for the LTO. The average total sales and the LTO mix was used fo estimate total LTO sales during the entire promotion. Manufacturers were asked fo manufacture product for 40% of the expected demand before the LTO was initiated. If was corroborated that competitors have similar practices. In fact, one competitor asked the manufacturer jo produce 100% of the expected sales in advance.
Since the forecast for a LTO was developed 10 weeks before inifiation, using the described forecasting process implied that the total weekly sales figure used together with the LTO mix was not “last week's” moving average. Manufacturers and distributors started procurement, manufacturing and logistics activities several weeks before orders from restaurants were observed. Thus, planning, and initial manufacturnng and inventory deployment were not based on “last week's” fotal sales, but ihe sales figure of the week before the memo was sent.
137
Actual Transactions from Information Systems
At the highest level of aggregation, the unit of analysis was the LTO. During a LTO a specially developed sandwich was offered in several menu items during a hmited time. For exampie, the promotional sandwich might be offered alone, in a combo, and in different meal sizes. The promotional meal was made of standard products and specifically developed ingredients for the LTO. The analysis in this research was focused on the ingredients particular to the promotional sandwich since the management of these products posed the challenges described in the section The Business Opportunity in Chapter 1.
The promotional sandwich included in this study used two specifically developed ingredients. Each of the two manufacturers supplied one of these two ingredients. The cistrioutor, who managed the nine distrioution centers, supplied the restaurants with not only the ingredients for ine promotional sandwiches, but other products. The ingredients for the promotional meals and other products flowed from the manufacturers to the distrioutors by the truckload and orders were placed once a week. The distributors replenished restaurants using milk-runs three times a week,
Both manufacturers provided data for all manufactured product for the LTO analyzed. However, only part of all manufactured products was shipped to the nine disirioution centers included in the analysis. Figure 4.4 is a surmmary of the data availability, The caill-outs represent percentages over the total product manufactured, Figure 4.4 snows that from all manufactured producti, 45% was shipped to the nine aistripbution centers included in the analysis. Point-of-sale data were available for 19% of ail restaurants of which slightly more than half were served by the distribution centers in the analysis. Additionally, point-of-sale data from some of these restaurants were incomplete. Restaurants with good data quality were manually selected. These restaurants represent 6.5% of the product manufactured.
138
All Restaurants _ 100%
Restaurants Point-of-Sale Data
FIGURE 4.4
Data Availability