制造商有海量的數據，但往往沒有正確的工具來開發它。這里有極大的潛力可挖。但如果你沒有這些工具，你該從哪里開始？遵循這六個步驟，開始從你的數據中獲得盡可能多的價值。

1. 數據整合

在制造業中，新傳感器采集的可用數據激增，而傳統的數據系統在處理和整合這些信息與現有來源方面存在困難。你的業務流程依賴于清楚、可靠的數據，從而帶來你在運營效率、客戶滿意度、財務業績等方面所期望的結果。
建立合適的基礎設施來協調和集中來自任何數量或源類型的數據，以確保在整個組織中使用通用定義，同時節省大量開發時間。

2. 數據治理

數據治理是成功的數據管理的一個主要組成部分。這是一個持續的過程，用于確定哪些數據對你的業務至關重要，并確保它保持正確的質量水平。關鍵是要為你的企業確定正確類型的治理框架，并定義員工需要遵循的流程。

生產、運營和業務對成功的看法都略有不同。你需要調整和管理你的數據，以確保他們目標一致。

3. 分析

數據可視化使你能夠以視覺上吸引人的格式瀏覽數據，并得出對企業成功至關重要的結論。通過從完全不同的來源獲取數據，對其進行轉換，并將其顯示在最終用戶可以看到和理解的儀表板中，你可以深入分析重要的KPI和指標。借助易于訪問的高級分析，找出差距和根本原因，并揭示趨勢。

4. 利益相關者權利

利益相關者的認同和持續支持對于數據項目的成功至關重要。確保自動化并在整個組織內分享見解，讓每個人隨時隨地都能看到事情的進展。

5. 變革管理

幾乎任何重大的技術或組織創新都需要對人們的工作方式做出同樣重大的改變。為了使項目成功并產生預期的價值，需要積極地管理組織變更。培訓、啟用和支持您的團隊，以確保你擁有合適角色的合適用戶，從而確保成功部署。

6. 演進

隨著你的不斷成長而發展！基于從第一步到第五步學到的知識進行迭代。

成果

你能期望從這樣的數據倡議中看到什么樣的結果？這里有幾個例子。
　
       · 結合生產力和財務數據，為生產經理顯示每條生產線的近乎實時的利潤產出，以幫助確定任何維護問題的優先級

· 將需求預測與生產計劃聯系起來，以確保供應得到優化，并確保正確的生產計劃到位，以限制低速SKU的過度生產

· 利用物聯網數據報告現場機器的健康狀況，主動降低維護成本，從而更好地分配現場技術人員

一旦你通過這些基本步驟建立了基礎，你就可以繼續探索高級分析和人工智能的可能性。

作者：Raz Nistor，Keyrus公司數據科學和CPG主任

文章原文：

6 Steps to Maximizing Value from Manufacturing Data

Manufacturers have tons of data but often don't have the right tools to explore it. There's a wealth of potential that's just waiting to be unleashed. But if you don’t have those tools in place, where do you start? Follow these six steps to start getting the most value possible from your data.

1. Data integration

In manufacturing, there’s an explosion of available data from new sensor sources, and legacy data systems struggle to process and combine this information with existing sources. Your business processes depend on clean, reliable data to produce the results you expect in terms of operational efficiency, customer satisfaction, financial performance, and more.

Set up the right infrastructure to harmonize and centralize your data from any number or type of sources to ensure that common definitions are used throughout the organization while saving significant development time.

2. Data governance

Data governance is a major component of successful data management. It’s a continuous process for identifying which data is critical to your business and ensuring it stays at the right level of quality. The key is to identify the right type of governance framework for your enterprise and to define the processes employees need to follow.

Production, operations, and the business all look at success slightly differently. You’ll need to align and govern your data to make sure they’re all looking at the same picture.

3. Analytics

Data visualizations allow you to explore your data in a visually appealing format and draw conclusions that are critical to the success of your business. By taking data from disparate sources, transforming it, and displaying it in dashboards where end users can see and understand it, you can drill in and analyze important KPIs and metrics. Find gaps and root causes, and uncover trends with easily accessible advanced analytics.

4. Stakeholder access

Stakeholder buy-in and continuous support are critical for data projects to succeed. Make sure to automate and share insights across the organization and allow everyone to see where things stand, any day, at all times.

5. Change management

Almost any significant technical or organizational initiative requires equally significant changes to the way people work. That organizational change needs to be actively managed in order for the project to be successful and generate the expected value. Train, enable, and support your team to ensure you have the right users in the right roles to ensure successful deployment.

6. Evolution

Evolve as you continue to grow! Iterate based on learnings from steps one through five.

Results

What kind of results can you expect to see from a data initiative like this? Here are a few examples.

Combined productivity and finance data to display the near real-time profit output of each line on the floor for production managers to help prioritize any maintenance issues
Connected demand forecasts with production schedules to ensure supply was optimized and that the right manufacturing schedules were in place to limit the overproduction of low-velocity SKUs
Proactively reduced maintenance costs using IoT data to report health of machines in the field, which leads to better allocation of field techs

Once you’ve laid the foundation with these basic steps, you can move on to exploring the art of the possible with advanced analytics and artificial intelligence.

About The Author
Raz Nistor is director of Data Science & CPG at Keyrus.