[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article:sansan-data-intelligence":3},{"meta":4,"markdown":187},{"type":5,"articleId":6,"slug":6,"title":7,"titleEn":8,"category":9,"summary":10,"publishedAt":11,"image":12,"vocabulary":13,"source":184},"article","sansan-data-intelligence","Sansan Data Intelligenceリリースに寄せて","On the Release of Sansan Data Intelligence","software\u002Fcompanies\u002Fsansan","Sansan announces Data Intelligence, a new product that manages enterprise master data by leveraging over 10 years of record-matching technology and a new time-series graph model called SOC v2.","2026-04-26T00:00:00Z","https:\u002F\u002Fimages.yamiyomi.com\u002Fsansan-data-intelligence.png",[14,19,24,28,32,36,40,44,48,52,56,60,64,68,72,76,80,84,88,92,96,100,104,108,112,116,120,124,128,132,136,140,144,148,152,156,160,164,168,172,176,180],{"word":15,"reading":16,"meaning":17,"level":18},"新規","しんき","new, fresh","N2",{"word":20,"reading":21,"meaning":22,"level":23},"意思決定","いしけってい","decision-making","N1",{"word":25,"reading":26,"meaning":27,"level":18},"進化","しんか","evolution",{"word":29,"reading":30,"meaning":31,"level":18},"挑戦","ちょうせん","challenge",{"word":33,"reading":34,"meaning":35,"level":23},"構想","こうそう","concept, plan",{"word":37,"reading":38,"meaning":39,"level":18},"短期間","たんきかん","short period",{"word":41,"reading":42,"meaning":43,"level":18},"実現","じつげん","realization, achievement",{"word":45,"reading":46,"meaning":47,"level":23},"蓄積","ちくせき","accumulation",{"word":49,"reading":50,"meaning":51,"level":23},"名寄せ","なよせ","record matching, deduplication",{"word":53,"reading":54,"meaning":55,"level":18},"既存","きそん","existing, established",{"word":57,"reading":58,"meaning":59,"level":18},"重複","ちょうふく","duplication, overlap",{"word":61,"reading":62,"meaning":63,"level":18},"統合","とうごう","integration, consolidation",{"word":65,"reading":66,"meaning":67,"level":18},"顧客","こきゃく","customer, client",{"word":69,"reading":70,"meaning":71,"level":18},"基盤","きばん","foundation, infrastructure",{"word":73,"reading":74,"meaning":75,"level":18},"構築","こうちく","construction, building",{"word":77,"reading":78,"meaning":79,"level":18},"表記","ひょうき","notation, written form",{"word":81,"reading":82,"meaning":83,"level":18},"更新","こうしん","update, renewal",{"word":85,"reading":86,"meaning":87,"level":18},"精度","せいど","accuracy, precision",{"word":89,"reading":90,"meaning":91,"level":23},"属人化","ぞくじんか","person-dependent process",{"word":93,"reading":94,"meaning":95,"level":18},"営業","えいぎょう","sales, business",{"word":97,"reading":98,"meaning":99,"level":18},"放置","ほうち","neglect, leaving as-is",{"word":101,"reading":102,"meaning":103,"level":18},"合併","がっぺい","merger",{"word":105,"reading":106,"meaning":107,"level":18},"請求書","せいきゅうしょ","invoice",{"word":109,"reading":110,"meaning":111,"level":23},"与信","よしん","credit assessment",{"word":113,"reading":114,"meaning":115,"level":23},"実害","じつがい","actual harm, real damage",{"word":117,"reading":118,"meaning":119,"level":23},"傘下","さんか","under the umbrella of",{"word":121,"reading":122,"meaning":123,"level":23},"継承","けいしょう","inheritance, succession",{"word":125,"reading":126,"meaning":127,"level":23},"昇華","しょうか","sublimation, elevation",{"word":129,"reading":130,"meaning":131,"level":18},"高度","こうど","advanced, sophisticated",{"word":133,"reading":134,"meaning":135,"level":18},"従来","じゅうらい","conventional, traditional",{"word":137,"reading":138,"meaning":139,"level":18},"構造","こうぞう","structure",{"word":141,"reading":142,"meaning":143,"level":18},"採用","さいよう","adoption, employment",{"word":145,"reading":146,"meaning":147,"level":23},"時系列","じけいれつ","time-series, chronological",{"word":149,"reading":150,"meaning":151,"level":18},"連続性","れんぞくせい","continuity",{"word":153,"reading":154,"meaning":155,"level":23},"断絶","だんぜつ","severance, discontinuity",{"word":157,"reading":158,"meaning":159,"level":18},"直面","ちょくめん","confrontation, facing",{"word":161,"reading":162,"meaning":163,"level":18},"有効","ゆうこう","valid, effective",{"word":165,"reading":166,"meaning":167,"level":18},"破壊","はかい","destruction",{"word":169,"reading":170,"meaning":171,"level":18},"保持","ほじ","retention, maintaining",{"word":173,"reading":174,"meaning":175,"level":18},"証明","しょうめい","proof, verification",{"word":177,"reading":178,"meaning":179,"level":23},"礎","いしずえ","cornerstone, foundation",{"word":181,"reading":182,"meaning":183,"level":18},"記述","きじゅつ","description, notation",{"name":185,"url":186},"Sansan Builders Box","https:\u002F\u002Fbuildersbox.corp-sansan.com\u002Fentry\u002F2025\u002F12\u002F12\u002F163000","\n::para\nSansan Data IntelligenceのProduct Owner（PO）を[務めて]{つとめて:serving as:N3}います[猿田]{さるた:Saruta (surname):N1}[貴之]{たかゆき:Takayuki (given name):N1}と[申し]{もうし:to be called:N3}ます。[本日]{ほんじつ:today:N5}12[月]{がつ:month:N5}12[日]{にち:day:N5}、[私]{わたし:I:N4}たちは[約]{やく:approximately:N3}4[年]{ねん:year:N5}ぶりとなる[新規]{しんき:new:N3}プロダクトSansan Data Intelligenceをリリースいたしました。これは、Sansanが[名刺]{めいし:business card:N2}[管理]{かんり:management:N2}サービスから、[企業]{きぎょう:company:N1}の[意思]{いし:intention:N4}[決定]{けってい:decision:N3}を[支える]{ささえる:to support:N3}ビジネスインフラへと[進化]{しんか:evolution:N3}するための[挑戦]{ちょうせん:challenge:N1}の[一つ]{ひとつ:one:N5}になります。\n\n#en\nMy name is Takayuki Saruta, and I serve as the Product Owner (PO) of Sansan Data Intelligence. Today, December 12th, we released Sansan Data Intelligence — our first new product in about four years. This represents one of our challenges as Sansan evolves from a business card management service into business infrastructure that supports corporate decision-making.\n::\n\n::heading\n[背景]{はいけい:background:N3}：[名寄せ]{なよせ:record matching:N3}[技術]{ぎじゅつ:technology:N2}の[蓄積]{ちくせき:accumulation:N1}\n\n#en\nBackground: Accumulation of Record-Matching Technology\n::\n\n::para\n[今回]{こんかい:this time:N3}のリリースは[構想]{こうそう:concept, plan:N3}から[約]{やく:approximately:N3}[半年]{はんとし:half a year:N5}という[短期間]{たんきかん:short period:N2}で[実現]{じつげん:realization:N3}しましたが、その[背景]{はいけい:background:N3}には10[年]{ねん:year:N5}[以上]{いじょう:or more:N4}にわたる[名寄せ]{なよせ:record matching:N3}[技術]{ぎじゅつ:technology:N2}の[蓄積]{ちくせき:accumulation:N1}と、[既存]{きそん:existing:N1}のデータベースアーキテクチャをリビルドするSOC v2の[開発]{かいはつ:development:N4}がありました。[名寄せ]{なよせ:record matching:N3}とは、[複数]{ふくすう:multiple:N2}のデータベースに[登録]{とうろく:registered:N2}されている[顧客]{こきゃく:customer:N1}[情報]{じょうほう:information:N3}から、[重複]{ちょうふく:duplication:N2}する[部分]{ぶぶん:part:N3}を[洗い出し]{あらいだし:identifying:N3}、[一つ]{ひとつ:one:N5}に[統合]{とうごう:integration:N1}する[作業]{さぎょう:work, task:N4}を[意味]{いみ:meaning:N4}します。\n\n#en\nThis release was achieved in just about half a year from conception, but behind it lies over 10 years of accumulated record-matching technology and the development of SOC v2, which rebuilds the existing database architecture. Record matching (nayose) refers to the process of identifying duplicates across multiple databases of customer information and consolidating them into one.\n::\n\n::heading\nなぜ[今]{いま:now:N5}、Sansan Data Intelligenceなのか？\n\n#en\nWhy Sansan Data Intelligence Now?\n::\n\n::para\n[現在]{げんざい:currently:N3}、[多く]{おおく:many:N4}の[企業]{きぎょう:company:N1}が「2025[年]{ねん:year:N5}の[崖]{がけ:cliff:N1}」[対策]{たいさく:countermeasure:N1}や[生成]{せいせい:generative:N3}AI[活用]{かつよう:utilization:N3}に[向けた]{むけた:aimed at:N3}データ[基盤]{きばん:foundation:N1}[構築]{こうちく:construction:N2}を[進めて]{すすめて:advancing:N3}います。しかし、[企業]{きぎょう:company:N1}の[約]{やく:approximately:N3}8[割]{わり:80 percent:N3}が「[取引先]{とりひきさき:business partner:N3}データに[重複]{ちょうふく:duplication:N2}・[表記]{ひょうき:notation:N3}[揺れ]{ゆれ:variation:N1}・[更新]{こうしん:update:N3}[漏れ]{もれ:omission:N1}などの[経験]{けいけん:experience:N3}がある」と[回答]{かいとう:respond:N3}しており、AI[活用]{かつよう:utilization:N3}に[取り組む]{とりくむ:to tackle:N3}[企業]{きぎょう:company:N1}の[約]{やく:approximately:N3}9[割]{わり:90 percent:N3}が「[期待]{きたい:expectation:N3}[通り]{どおり:as expected:N4}の[精度]{せいど:accuracy:N3}が[出ない]{でない:does not come out:N5}」と[嘆いて]{なげいて:lamenting:N1}います。いわゆる「Garbage In, Garbage Out」の[状況]{じょうきょう:situation:N2}です。\n\n#en\nCurrently, many companies are building data foundations for \"2025 cliff\" countermeasures and generative AI adoption. However, about 80% of companies report experiencing duplication, notation variations, and update omissions in their business partner data, and about 90% of companies working on AI adoption lament that they cannot achieve expected accuracy. This is the classic \"Garbage In, Garbage Out\" situation.\n::\n\n::heading\n[現場]{げんば:on-site:N3}で[起きて]{おきて:occurring:N4}いる[負]{ふ:negative:N3}のループ\n\n#en\nThe Negative Loop Happening on the Ground\n::\n\n::para\n[入力]{にゅうりょく:input:N4}の[属人化]{ぞくじんか:person-dependent:N1}：[現場]{げんば:on-site:N3}の[営業]{えいぎょう:sales:N2}[担当者]{たんとうしゃ:person in charge:N2}が[手]{て:hand:N4}[入力]{にゅうりょく:input:N4}するため、[社名]{しゃめい:company name:N4}が「（[株]{かぶ:stock\u002FInc.:N1}）」だったり「[株式]{かぶしき:stock:N1}[会社]{かいしゃ:company:N4}」だったりバラバラになる。[更新]{こうしん:update:N3}の[放置]{ほうち:neglect:N3}：[移転]{いてん:relocation:N2}や[合併]{がっぺい:merger:N1}があっても、[古い]{ふるい:old:N4}[情報]{じょうほう:information:N3}がそのまま[残る]{のこる:to remain:N3}。[結果]{けっか:result:N1}、[請求書]{せいきゅうしょ:invoice:N1}が[届かない]{とどかない:does not arrive:N2}、[与信]{よしん:credit:N3}[管理]{かんり:management:N2}が[漏れる]{もれる:to slip through:N1}といった[実害]{じつがい:actual harm:N3}が[出る]{でる:to come out:N5}。\n\n#en\nPerson-dependent input: Because on-site sales staff enter data manually, company names end up inconsistent — sometimes \"(Inc.)\" and sometimes \"Corporation.\" Neglected updates: Even when relocations or mergers occur, old information remains as-is. As a result, real harm occurs — invoices don't arrive, credit management falls through the cracks.\n::\n\n::heading\nData Hubとの[関係性]{かんけいせい:relationship:N3}\n\n#en\nRelationship with Data Hub\n::\n\n::para\nData Hubでも[高]{こう:high:N5}[精度]{せいど:accuracy:N3}な[名寄せ]{なよせ:record matching:N3}は[可能]{かのう:possible:N3}ですが、「A[社]{しゃ:company:N4}とB[社]{しゃ:company:N4}が[合併]{がっぺい:merger:N1}してC[社]{しゃ:company:N4}になった」「D[社]{しゃ:company:N4}はEグループの[傘下]{さんか:under the umbrella of:N1}に[入った]{はいった:entered:N5}」といった[複雑]{ふくざつ:complex:N2}な[企業間]{きぎょうかん:inter-company:N1}の[関係性]{かんけいせい:relationship:N3}を[表現]{ひょうげん:express:N3}して、[管理]{かんり:management:N2}することには[限界]{げんかい:limit:N3}がありました。つまり、Data Hubで[培った]{つちかった:cultivated:N1}[名寄せ]{なよせ:record matching:N3}[技術]{ぎじゅつ:technology:N2}という[資産]{しさん:asset:N3}を[継承]{けいしょう:inheriting:N1}しつつ、それを[企業]{きぎょう:company:N1}[全体]{ぜんたい:entire:N3}のデータガバナンス[基盤]{きばん:foundation:N1}へと[昇華]{しょうか:sublimation, elevation:N1}させたのがSansan Data Intelligenceです。\n\n#en\nWhile high-accuracy record matching was possible with Data Hub, there were limits to expressing and managing complex inter-company relationships like \"Company A and B merged to become Company C\" or \"Company D came under the E Group umbrella.\" In other words, Sansan Data Intelligence inherits the record-matching technology asset cultivated in Data Hub while elevating it into a data governance foundation for the entire enterprise.\n::\n\n::heading\nSOC v2の[実現]{じつげん:realization:N3}\n\n#en\nRealizing SOC v2\n::\n\n::para\nこの[高度]{こうど:advanced:N4}なデータマネジメントを[実現]{じつげん:realization:N3}している[裏側]{うらがわ:behind the scenes:N2}には、SOC v2 (Sansan Organization Code version 2)というコードがあります。[従来]{じゅうらい:conventional:N1}のSOCシステムではRDBのテーブル[構造]{こうぞう:structure:N2}を[採用]{さいよう:adoption:N2}していましたが、[時系列]{じけいれつ:time-series:N1}[方向]{ほうこう:direction:N3}に[遡れない]{さかのぼれない:cannot trace back:N1}、[合併]{がっぺい:merger:N1}で[新]{しん:new:N4}[会社]{かいしゃ:company:N4}ができるとIDが[変わり]{かわり:changes:N3}、[過去]{かこ:past:N3}のIDとの[連続性]{れんぞくせい:continuity:N3}が[断絶]{だんぜつ:severed:N3}してしまうという[技術的]{ぎじゅつてき:technical:N2}[課題]{かだい:issue:N2}に[直面]{ちょくめん:confronting:N3}していました。\n\n#en\nBehind this advanced data management lies a code called SOC v2 (Sansan Organization Code version 2). The conventional SOC system used an RDB table structure, but it faced technical challenges: it could not trace back through time-series data, and when mergers created new companies, IDs would change and continuity with past IDs was severed.\n::\n\n::heading\n[時系列]{じけいれつ:time-series:N1}グラフモデル\n\n#en\nTime-Series Graph Model\n::\n\n::para\nこれらを[解決]{かいけつ:resolution:N3}するために、SOC v2は[時系列]{じけいれつ:time-series:N1}グラフモデルを[採用]{さいよう:adoption:N2}しました。[最大]{さいだい:greatest:N3}の[特徴]{とくちょう:feature:N1}は、すべてのEdgeが[有効]{ゆうこう:valid:N2}[期間]{きかん:period:N3}を[持って]{もって:having:N4}いることです。[移転]{いてん:relocation:N2}や[合併]{がっぺい:merger:N1}などの[変更]{へんこう:change:N3}は、レコードの[上書き]{うわがき:overwrite:N5}ではなく、[新しい]{あたらしい:new:N4}Edgeを[追加]{ついか:addition:N3}する[更新]{こうしん:update:N3}で[表現]{ひょうげん:express:N3}されます。これによって、[過去]{かこ:past:N3}のいかなる[時点]{じてん:point in time:N3}の[状態]{じょうたい:state:N1}も[破壊]{はかい:destruction:N1}することなく[保持]{ほじ:retention:N1}し[続ける]{つづける:to continue:N3}ことができます。\n\n#en\nTo solve these issues, SOC v2 adopted a time-series graph model. Its greatest feature is that all Edges have a validity period. Changes such as relocations and mergers are expressed not by overwriting records but by adding new Edges. This makes it possible to retain the state at any past point in time without destruction.\n::\n\n::heading\nおわりに\n\n#en\nConclusion\n::\n\n::para\nSansan Data IntelligenceとSOC v2は、[単なる]{たんなる:mere:N3}マスターデータ[管理]{かんり:management:N2}システムではありません。それは、「いつ、[誰]{だれ:who:N3}が、どのような[状態]{じょうたい:state:N1}で[存在]{そんざい:existence:N3}したか」を[証明]{しょうめい:proof:N1}するデータ・ガバナンスのインフラです。Data Hubで[培った]{つちかった:cultivated:N1}[技術]{ぎじゅつ:technology:N2}を[礎]{いしずえ:cornerstone:N1}に、この[新規]{しんき:new:N3}プロダクトを[得て]{えて:gaining:N3}、Sansanは[世界]{せかい:world:N4}のビジネス[構造]{こうぞう:structure:N2}そのものを[記述]{きじゅつ:describe:N2}するプラットフォームへと[進化]{しんか:evolution:N3}します。\n\n#en\nSansan Data Intelligence and SOC v2 are not mere master data management systems. They are data governance infrastructure that proves \"when, who, and in what state something existed.\" With the technology cultivated in Data Hub as a cornerstone, and with this new product, Sansan evolves into a platform that describes the very structure of global business.\n::\n"]