百分點數據科學實驗室：菸草行業市場信息採集數據質量評估體系研究探索

{"type":"doc","content":[{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"編者按","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"對於菸草企業來說，掌握真實市場消費信息對於貨源的精準投放、研究市場的真實需求、完善區域內捲菸產品進退的機制、探索捲菸產品發展週期的拐點等起着重要的參考作用。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"未來，在捲菸市場全面進入數字化時代後，通過採集市場信息所形成的數據分析將在上述業務環節中起到關鍵性、決定性的作用，因此如何有效地針對市場信息的數據觸點（雲pos終端機）建立起數據質量評估體系，科學引導零售戶正確使用終端機，從而提高整體數據質量，獲取真正實用的基礎數據，是當前菸草企業首要解決的核心問題。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在本篇文章中，百分點數據科學實驗室基於實際項目中的實踐經驗，總結了一套菸草行業市場信息採集、數據質量評估體系建設的方法論。","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"一、相關概念介紹","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"終端機：","attrs":{}},{"type":"text","text":"指在門店部署雲Pos終端機的零售戶門店，由雲pos終端機提供市場信息採集所需要的數據，所有配套了雲pos終端機的零售戶門店統稱爲雲pos終端（以下統稱終端機），而配套了終端機的零售戶實現了捲菸或非煙商品在進、銷、存方面的自動化數據管理，可爲市場信息採集提供時效性較高的數據資源，未來終端機將逐步全面覆蓋所有的零售戶，爲數字化捲菸銷售提供硬件基礎。","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/27/2787cca243a85aec926ce03057bd6487.jpeg","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"終端機數據質量：","attrs":{}},{"type":"text","text":"指對終端機提供的進銷存數據在真實性、完整性、準確性的要求，數據被評估與判斷爲質量越好的終端，其所提供的市場消費數據可深入分析的價值越大，可在捲菸市場價格指數等主題分析的探索研究中起到重要作用。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"市場分析樣本：","attrs":{}},{"type":"text","text":"基於樣本估計總體的分析思路，根據層級結構、數據質量、樣本庫數量要求抽取符合要求的市場信息樣本點進行數據分析，樣本點由信息採集點中擇優選取，存在一定的比例要求。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/83/8327aba65f3007b2a6bdeb7eff5866ad.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/c7/c7175dea4eac77bd08eb1f2923bd0e49.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"二、市場採集的價值","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"洞察市場 優化捲菸經營環境","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在營銷轉型的新形勢下，菸草行業對於社會消費、庫存數據的分析與應用具有迫切性。從微觀層面看，菸草行業迫切需要了解每一款投放產品真實的市場狀態，準確識別產品所處的生命週期階段，爲捲菸產品在各區域的進退提供數據依據；從宏觀層面看，迫切需要對捲菸市場的社會庫存壓力進行科學評估，準確把握捲菸發展拐點，在工商協議的簽訂、貨源投放的精準、社會庫存結構的優化方面向數據驅動模式轉變。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"同時，在捲菸專賣管理數據分析方面，市場信息採集對市場經營環境起到淨化作用，如對社會庫存的研究，可從源頭上發現真煙異常流動（區域間竄貨）的違規行爲，提前響應市場管理機制，進一步加強對捲菸市場的管理水平。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"綜上所述，市場信息採集對於捲菸行業的重要性逐漸突顯，當前市場信息採集的獲取途徑是零售戶門店所部署的雲pos終端機，它們構成了市場採集的數據觸點網絡，是實現自動化信息採集的硬件基礎。","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"發展模式的轉變","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"基於市場採集所反饋的數據，菸草企業能夠掌握更多的市場真實消費信息，加深對消費者行爲的研究，不斷挖掘出市場的真實需求，在有效幫助零售戶提高盈利水平的同時，還能不斷增強菸草企業的渠道掌控力，實現由傳統捲菸批發商向以數據驅動主動力的模式轉變，加快企業由捲菸批發企業向新零售供應商的升級轉變。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/40/40bee16472465e58e4ceee59925e70e9.jpeg","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"三、市場採集的問題","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"1. 數據質量要求","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"根據需求調研，從數據可應用的角度來講，市場採集數據只有合格與不合格兩種，主要體現在三個方面：","attrs":{}}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"真實性，即所採集的數據能真實反映終端的銷售情況，否則採集就失去意義；","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"準確性，即所採集的數據準確無誤，不能有偏差，否則不能反映真實的市場情況；","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"完整性，即除了捲菸商品外，還需要採集非煙商品的銷售數據，才能更準確的反映出該終端所面臨的消費羣特徵。","attrs":{}}]}]}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/10/10a5cbebacdec553cc90e70f679c1092.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"只有符合真實性、準確性、完整性三個特點的數據纔是合格數據。當前，造成數據不合格的主觀因素與客觀因素如下：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"主觀因素","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"軟件使用者在主觀上不想把真實的銷售情況通過掃碼銷售反映出來，因此存在集中掃碼、數據修改頻繁、庫存準確率低等問題。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"客觀因素","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"即使軟件使用者願意將真實銷售情況通過掃碼銷售上傳到菸草公司，但由於使用者能力不足、指導人員沒有進行有效指導、設備故障等，造成數據不準確、不完整，導致數據準確率低、非煙商品掃碼數量與種類低、庫存低於合理值等。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"根據實際業務調研，目前市場採集尚處於終端機設備的初始推廣階段，零售戶在主觀與客觀方面的因素造成了原始數據質量不高，因此需要在初始推廣階段，結合客戶經理基層拜訪的實際工作情況，形成適合當前階段的零售戶終端機使用行爲評估與樣本點篩選體系，基於體系化管理引導零售戶羣體規範化終端設備的使用行爲，逐步將整體數據質量提升，爲深層次的主題分析打下數據基礎。","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"2. 質量評估量化","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"進銷庫存數據主要由4張基礎表組成：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"登陸日誌表","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/d0/d0a7eeb590f5ec0cbfc9a19beb7fc6d3.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"銷售單行表","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/c1/c17f662c5388939b9e0611ec1e937af4.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"入庫單行表","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/7a/7a37c3a1981588a234c50950f9981ddc.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"客戶商品庫存表","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"基於終端機的數據源表結構，結合對實際業務的調研與總結，本文初步探索出以下的可量化分析維度：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"終端掃碼時段","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"終端掃碼時段是指每日終端機存在有掃碼動作記錄的時段，一天共有24個時段，若零售戶在1天的有效經營行爲中，銷售執行動作的掃碼時段較少，則屬於不正常掃碼現象，需要及時介入引導。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"系統在線時長","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"每個終端機都通過有線網絡連接至後端服務器，且均有配套的管理賬戶，當零售戶登錄終端機系統後，後臺服務器將持續記錄其登錄狀態，若終端機斷網，則後端服務器顯示終端機賬戶離線，可能存在經營異常，需要及時反饋實際經營情況。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"掃碼金額、零售量","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"掃碼金額，指單個商品或者單筆的銷售總金額、總銷售量，若出現較大的異常值，則可能存在集中刷、亂刷的使用行爲，目前僅限於對捲菸商品的監控。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"掃碼筆數","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"掃碼筆數，指銷售的訂單數量，一筆訂單包含多個商品，一筆掃碼代表一筆訂單，若出現較大的異常筆數，則可能存在集中刷、亂刷的使用行爲。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"掃碼品規","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"掃碼品規，指每筆訂單中包含捲菸品規的數量，若單筆訂單中出現較多的品規，SKU寬度較大，則可能存在集中刷行爲。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"在線支付","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在線支付，指的是消費者主動出示二維碼支付的行爲，若當日在線支付比例較低，則說明零售戶存在不真實掃碼的情況。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"庫存量","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"庫存量，指的是商品發生銷售行爲後、商品消耗後存儲量的變化情況，若商品出現負庫存的情況，則可能存在超量銷售的情況，目前僅限於對捲菸商品的監控。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"集中度","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"集中度，指的是單日掃碼筆數最高的兩個時段掃碼筆數之和/當日總掃碼筆數，目前僅限於對捲菸商品的監控。","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"3. 樣本結構問題","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"根據實際的業務要求，目前市場的數據採集範圍僅限於樣本點，總體思路爲通過樣本點數據估計總體趨勢，原則上樣本點的結構要求做到各行政區域內檔位、業態、城鄉的全覆蓋，且要求符合各層級（檔位+業態+城鄉）實際分佈，但不同的層級具有不同的經營能力，各層級所需抽樣的數量還需要根據各層級的經營能力判斷，層級結構如下:","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"區域","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"行政區劃分，如芙蓉區、天心區等。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"檔位","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"共有30個等級，主要用於衡量零售戶的經營水平，級別越高，經營水平越強。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"業態","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"區分零售戶的經營業態，主要由超市、菸酒店、便利店、食雜店等組成，相同檔位內不同業態的經營能力也不一樣，例如同檔位內超市與菸酒店的經營能力有顯著的差異。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"城鄉","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"區分零售戶所處區域特徵，共有2個屬性維度，城市與鄉村，城鄉屬性的零售戶經營能力普遍高於鄉村屬性。","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"4. 樣本庫管理問題","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"隨着終端質量的不斷提高，未來樣本庫的更新需要具備一套自動化的新老輪換機制，來實現對樣本庫的動態管理，對於無法達到樣本點要求的老採集點，原則上要執行置換過程，置換的過程需要有量化支持，有理可依、有數可查。置換原則通過業務探索，總結如下：","attrs":{}}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"新老採集點置換時必須滿足同類別置換的原則；","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"新採集點客戶需配合程度較好，便於管理；","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"新採集點需儘量避開偏遠地區和文化水平不高的客戶；","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"新採集點需經營穩定，短期內無需再次置換；","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"新採集點不要選擇近期可能出現較大檔位變化的客戶；","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"樣本置換須報信息管理員做好備案工作。","attrs":{}}]}]}],"attrs":{}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"四、解決方案","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"基於3σ 模型與專家業務決策的數據質量分析體系","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"針對終端機數據質量評估問題，在統計學方面本質上屬於對離羣值的判斷。在終端機掃碼銷售的過程中，大部分出現異常的掃碼行爲均可量化與轉化爲統計學上對離羣值的判斷。總結爲基於概率分佈模型進行離羣值的判斷，其基本思想是根據數據集的特性，對給定的數據集假設一個概率分佈模型，然後在某個顯著水平上，確定數據集的拒絕域或者接受域，拒絕域是出現概率很小的區間，如果數據落在此區間，則可判定爲異常數據。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"本次研究中，在假設整體屬於正態分佈的情況下，對指標規則的制定主要應用3倍均值標準差（3σ）進行離羣值的判斷。根據大數定律和中心極限定理可知，當樣本數據足夠多時，其分佈趨於正態分佈。在正態分佈的基礎上，常用異常值被定義爲一組結果值中與平均值的偏差超過三倍標準差的值，理論分佈圖如下圖所示。","attrs":{}}]},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/b7/b7883eae8b8492d449894e7f069524de.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"除了在統計學層面外，在設計終端數據質量評估體系的過程中，更應該注意專家業務決策的經驗性指導，不可忽視其關鍵作用。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"總結以上方法論，項目實施過程中我們提出了日常監控 + 月度評分與分類的終端數據質量評估體系，其中日常監控的作用在於：","attrs":{}}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"觀察零售戶在使用終端機的過程中具體會有哪些異常的掃碼行爲，在數據層面挖掘零售戶羣體在使用終端的過程中造成數據質量低下的核心原因，制定宏觀策略；","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"輔助客戶經理在市場走訪過程中針對性地進行終端機業務指導，對每日產生的終端使用預警進行排查，引導零售戶正確使用終端機，逐步將“數據池”清洗乾淨；","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"爲月度樣本點的評估提供日度數據參考；","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"實現指標參數環境自適應調節機制，不同區域、層級各自對應的規則參數均不同，形成差異化數據質量判斷。","attrs":{}}]}]}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在日常監控環節，項目主要形成的規則體系如下：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/e5/e5595347fd684a02ffd336aa0f784d63.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":"center","origin":null},"content":[{"type":"text","text":"（注：同羣體 = 同層級）","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"樣本點分層隨機抽樣法之“奈曼分配”","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"針對區域+檔位+業態+城鄉組成的樣本點層級結構要求，經過項目組分析，其理論上屬於分層抽樣的範疇，分層抽樣又稱分類抽樣或類型抽樣，將總體劃分爲若干個同質層，再在各層內隨機抽樣或機械抽樣。分層抽樣的特點是將科學分組法與抽樣法結合在一起，分組減小了各抽樣層變異性的影響，抽樣保證了所抽取的樣本具有足夠的代表性，各層樣本數的確定方法有3種：","attrs":{}}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"分層定比法，即各層樣本數與該層總體數的比值相等，例如，樣本大小n=50，總體N=500，則n/N=0.1即爲樣本比例，每層均按這個比例確定該層樣本數；","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"奈曼法，即各層應抽樣本數與該層總體數及其標準差的積成正比；","attrs":{}}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"非比例分配法，當某個層次包含的個案數在總體中所佔比例太小時，爲使該層的特徵在樣本中得到足夠的反映，可人爲地適當增加該層樣本數在總體樣本中的比例。","attrs":{}}]}]}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/fa/fa354a10e38c93bbf17cc85bb97eda7c.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"因各個層級之間的差異性較大，主要體現在捲菸訂單（進貨）數量方面，若要較爲真實地接近各層級的經營水平，需要引入訂單（進貨）因素，通過結合奈曼分配法，基於層級羣體的訂單（進貨）量的標準差與客戶數確定每層比例，最後根據所制定樣本量合理分配各層級中抽樣數量，項目實施過程中，還引入了終端數據質量評分體系，在奈曼分配法的基礎上，按數據質量分數的高低擇優抽樣入庫，樣本點篩選邏輯如下：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/8f/8faaa7ef41c9db7434a46f47ddd7dd1e.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"樣本庫量化管理機制","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"針對樣本庫量化管理方面的需求，本文采用系統日常評價+月度質量評估的機制對所有在線正常運行的雲pos終端機進行量化評分，評分的高低結合奈曼法進行樣本點的動態管理。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/5f/5fdbede312872aa0649733d5b6de38c4.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在日常監控評估環節，本文采用滿分扣分制度，每日初始滿分爲100分，若當日觸發不同的監控規則扣除規則對應的分數，分數扣除越多，則當日的數據越低，分數越低則影響日均評分成績。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在月度量化評分環節，則是根據項目要求，按月的時間維度對所有的終端機使用情況進行評分，分數越高，數據質量越好。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"實現對終端機“平時成績”與“月度期末考試”的評價體系搭建後，系統將在每月初通過綜合評定對終端機使用情況進行量化，其中，日常監控評分權重爲0.6，月度評估評分權重爲0.4，兩者綜合分數高者所劃分的羣體等級越高，等級越高代表終端使用質量越高，其數據利用價值越高，便越接近真實的市場銷售行爲。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在實際應用中，爲了加強樣本庫的穩定性，對一類、二類樣本點設置了容錯保護機制，具體體現爲基礎加分與進退規則的差異化，避免大規模樣本點置換現象出現。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"月度量化評分規則體系如下：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/f1/f1d8f79adb8dccf0d4934fc1c59b857d.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"等級劃分規則與等級進退原則如下：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/1d/1d80dd8a4ffc84bb214ebc2351531a54.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"五、實際應用成果","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"1. 應用結果","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"日常監控方面，規則部署完成系統化開發，實現了在PC端、移動端對零售戶終端機的規則監控，T-1日推送異常信息，幫助客戶經理及時介入。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"PC端應用成果：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/49/499da99105a7dd55e24e4601562dfa34.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"移動端應用成果：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"形成了監控→預警→推送→覈驗→反饋的業務閉環，爲客戶經理提供了移動辦公的條件。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/96/967debc786247957c2898a7f9bd9d8b8.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"數據質量提升效果：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"系統於2021年3月在某市局上線，通過現行的規則監控與線下終端引導，零售戶月度人均規則數逐月下降，特別是在5月推進“終端全優化”專項行動後，終端機配套數量不斷提升的背景下，人均命中規則數降低至今年最低點，說明零售戶整體的終端機掃碼質量在顯著提升。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/05/05641358a719018af26b1d7f5eade515.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/3c/3c907aab63cd4bce97a2ff7f3b323fab.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"樣本點遴選結果：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"根據樣本點佔總體1%的要求，實現了4個試點區域的樣本點的篩選工作，並實現了系統化的開發。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/0d/0dccc2c9b1eb2ac7b357f75883f248e8.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"PC端樣本點管理界面：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/1f/1fd167422d420aa684193731bccb9ae7.png","alt":"圖片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"2. 未來可研究方向","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"未來隨着市場信息採集數據質量評估體系的實施，菸草公司對於端數據的管理逐漸趨向規範化，整體終端的建設發展向高水準邁進，由此產生的數據將爲捲菸市場研究甚至是非煙商品的研究提供可靠的數據支持，如何將高質量的市場信息採集數據利用起來，是未來菸草行業進行深度市場化改革的探索道路之一，本文就以下四大研究方向進行了探索：","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"2.1 以採集端爲基礎建設大數據營銷體系","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在相關理論研究中，構建捲菸市場大數據營銷體系共需包括三個平臺，即採集、分析和應用平臺。採集端是分析與應用端的基礎，在完善採集端的建設後，由此基礎上可進行整合建立起包括數據採集系統、數據分析系統、應用系統的菸草大數據營銷體系。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"數據採集系統","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"企業通過建立詳盡的客戶基礎信息檔案，包括內部訂單與外部零售的數據資源，形成真正的捲菸大數據並進行數據儲備。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"數據分析系統","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"應用主體是捲菸企業，構建數據科學的分析模型，深入挖掘有價值的數據，提煉出對能夠促進市場、品牌、終端建設有價值的信息知識。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"數據應用系統","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"應用主體是企業營銷決策者，在數據科學所獲得的信息知識下進行營銷決策，營銷決策直達捲菸市場，營銷決策所帶來的市場效應將由採集系統供應數據、分析系統提供結果。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"以上三者之間形成一種相互聯繫、相互貫通的數據閉環關係，真正實現數據驅動模式的市場化改革。","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"2.2 以市場消費數據爲基礎進行市場供需關係預測","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"通過市場採集數據對市場狀態進行科學判斷，首先要搭建市場量化指標，如基於售價數據的指標：零售與市場價格指數、條包零售價吻合率等；基於庫存數據的指標：社會庫存、社會存銷比等；基於銷量數據的指標：零售量等。根據歷史經驗，價格指數與零售量提升、社會存銷比趨勢下降，就說明市場供需緊張、市場處於空缺狀態；反之，則說明貨源過剩、供需偏松、市場處於飽和狀態。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"同時，針對當前捲菸供應計劃與訂單需求脫節時會導致供需關係背離的情況，還可引入內部訂單數據中的捲菸訂足率作爲輔助指標，最後根據內外部數據，搭建較爲真實的市場需求預測模型，相對於純內部數據時間序列、僅結合外部宏觀經濟數據的迴歸模型等，其更能體現真實的市場需求、更具業務解釋性、對消費者市場的供需關係預測更爲科學準確。","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"2.3 以市場依據爲導向進行品牌生命週期管理","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在捲菸市場中，如何準確判斷品牌的生命週期、把握品牌發展拐點、準確制定品牌經營策略，是菸草企業一直探索的行業難題。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"根據以往資料研究，捲菸企業對於品牌生命週期的判斷基本來源於內部的訂單數據，對外部的市場消費數據的參考度較低，較容易忽略真實的品牌狀況，在未來以市場導向爲改革路線的捲菸行業中，市場消費數據對捲菸行業的發展將日益重要。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"菸草企業可根據終端數據建立起以市場依據爲導向的品牌生命週期管理機制，如依據某一時期內品牌掃碼銷售數據，確定品牌的上櫃率、銷售增長率，以此將品牌發展階段劃分爲導入、成長、穩定和衰退四個階段，根據當前所處的階段，具體分析該階段品牌的動銷率、重購率、斷貨率等二級指標，同時結合內部訂單中重需率確定品牌的健康指數，全面量化品牌的市場健康狀態，科學判斷品牌生命週期，爲品牌的培育、引入、退出提供依據。","attrs":{}}]},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"2.4 依託終端建設開展個性化客戶服務","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"對零售戶開展個性化精準服務是捲菸行業加強終端建設的重點方向，加強終端建設能夠有效提升渠道掌控力，對新時代菸草商業企業高質量發展具有重大意義。客戶經理可以利用平板電腦、手機等移動終端設備，根據後臺海量的終端數據隨時瞭解市場整體趨勢及終端零售客戶的基本情況，爲有針對性經營指導和拜訪提供了可能性。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"例如，在客戶經理拜訪時間管理方面，利用現代零售終端經營流水數據，分析出零售戶一天不同時段的經營繁忙程度，進而避開繁忙時段科學安排拜訪流程；在拜訪服務內容管理方面，通過整合自動識別的潛在服務需求、客戶直接服務需求等各種信息，將服務需求與營銷計劃進行匹配，自動生成個性化拜訪服務方案，從而在提高服務營銷方案制定效率的同時提高其針對性；在非煙商品推薦方面，依據關聯規則模型得出強關聯商品目錄，輔助指導零售戶在優化貨架擺放、銷售結構，提升銷售利潤率。此外，在完成個性化精準服務後，還可依託終端數據對服務水平進行量化評估，形成良性服務閉環。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"參考資料","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"[1]白明，推動捲菸零售終端轉型升級的探索研究，現代營銷（下旬刊），2019年4期.","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"[2]周亮，“大數據”技術在浙江菸草新時期的應用[J].硅谷,2013(19).","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"[3]劉洋，數據驅動創新大數據時代的捲菸營銷思考[J].信息與電腦(理論版)2013，（7）.","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"[4]鄒亮，捲菸零售終端價值評價體系研究，《中國菸草學報》，2013年4期。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"[5]許譯文，捲菸零售終端建設理論模型初探[J].新西部.2015(12):69.","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"[6]郭飛，構建複合型捲菸零售終端的路徑選擇[J].重慶與世界,2016(11):81-83.","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"[7]劉金娣，李莉莉，高靜，盧睿，異常值檢驗方法的比較分析[J].青島大學學報（自然科學版）第30卷第2期2017年5月.","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"原文鏈接：","attrs":{}},{"type":"link","attrs":{"href":"https://mp.weixin.qq.com/s/aESP1L204GSfR2Oy3Ar_tA","title":"","type":null},"content":[{"type":"text","text":"百分點數據科學實驗室：菸草行業市場信息採集數據質量評估體系研究探索","attrs":{}}]}]}]}