A solid foundation to building a scalable People Analytics function. You’ve got to start somewhere, you might as well start here.

建立可扩展的人员分析功能的坚实基础。 您必须从某个地方开始，不妨从这里开始。

People Analytics (also known as HR Analytics, Talent Analytics, and Workforce Analytics) is essentially the utilization of employee-related data to drive business decisions.

人员分析(也称为HR分析，人才分析和劳动力分析)实质上是利用与员工相关的数据来制定业务决策。

And it’s all the rage.

这就是所有的愤怒。

Data is everywhere, and every second, every mili-second, we are collecting more of it. Business leaders have learned that harnessing their data to make business decisions can have huge implications for their bottom line.

数据无处不在，每一秒钟，每毫秒，我们都在收集更多数据。 企业领导者了解到，利用他们的数据来制定业务决策可能对其底线产生巨大影响。

What types of ad campaigns generate the most sales? What kinds of products have the highest profit margins? Where does our target demographic do most of their shopping?

哪些类型的广告系列产生最多的销售额？ 哪种产品的利润率最高？ 我们的目标人群在哪里购物最多？

Collecting data, analyzing it, and then making decisions based off of those insights has become an expectation. And as business leaders have grown increasingly data savvy, there is a hot demand for ways to apply these principals to employee-related decisions. In comes People Analytics.

收集数据，对其进行分析，然后根据这些见解做出决策已成为一种期望。 随着企业领导者越来越多地掌握数据，人们迫切需要将这些原则应用于与员工相关的决策的方法。 引入人员分析。

If you want to see super cool examples, Google’s People Analytics team has made some of their most impactful projects openly available.

如果您想查看超酷的示例，则Google的People Analytics团队已公开开放了其一些最具影响力的项目。

Okay, so you’re bought in. You are convinced that People Analytics is going to change the world, make your business a million times more efficient, and enable your employees to be the most engaged employees on the planet. You want your organization to reap the benefits of this new field. Welcome to the cool kid’s table.

好的，您就这么买了。您确信People Analytics将改变世界，使您的业务效率提高一百万倍，并使您的员工成为地球上最投入的员工。 您希望您的组织从这个新领域中受益。 欢迎来到酷孩子的桌子。

But where do you start?

但你从哪儿开始呢？

Here are three things to keep in mind when first starting out that will help set up your People Analytics function with a solid foundation and gain traction quickly:

刚开始时，请牢记以下三件事，这将有助于您建立人脉分析功能并打下坚实基础：

1. Make it accurate

   使其准确

2. Make it actionable

   使其可行

3. Make it accessible

   使其可访问

1.使其准确(1. Make it accurate)

In an ideal world, all of your data will live together in a centrally managed database. As we all know, the world we live in is far from ideal and our data often lives in all kinds of places, comes from all kinds of sources, and is in all kinds of formats. Oh, the joys.

在理想的世界中，所有数据将一起存储在一个集中管理的数据库中。 众所周知，我们所生活的世界远非理想之境，我们的数据经常生活在各种地方，来自各种来源，格式各异。 哦，很高兴。

Does this make our job complicated? Yes. Does it make our job pull-our-hair-out frustrating at times? Also, yes. But there are a couple of things that you can establish early on that will save you A LOT of headache in the future:

这会使我们的工作复杂吗？ 是。 这是否会使我们的工作有时变得令人沮丧？ 还可以但是，您可以在早期建立一些条件，以免将来麻烦很多：

• Establish data ownership建立数据所有权
• Have a single source of truth有单一的真理来源
• Establish credibility early尽早建立信誉

建立数据所有权(如果需要，可以建立“管理权”)。(Establish data ownership (or “stewardship” if you want to be fancy).)

There are many ways to do establish data ownership to meet the unique needs of your organization. But for starters, you need to establish: who is responsible for what data, where does it live, how do you keep it secure and ensure it is accurate.

有很多方法可以建立数据所有权以满足组织的独特需求。 但是对于初学者来说，你需要建立：谁负责什么样的数据，它在哪里住，你怎么保持它的安全，并确保其准确。

Ha, that’s all?!

哈，仅此而已？

Again, this can seem overwhelming but stick with me intrepid warrior. Let’s try an example. Let’s say compensation data is processed by the Finance Department and is recorded in our Human Resources Information System (HRIS).

再说一次，这看起来似乎是压倒性的，但是我还是一个勇敢的战士。 让我们尝试一个例子。 假设薪酬数据是由财务部门处理的，并记录在我们的人力资源信息系统(HRIS)中。

We will say that each month Gracie Lou Freebush in Finance will run a report in our HRIS system and check it against signed offer letters from recruiting. She will encrypt the file and save it to our internal server.

我们会说，财务部门的Gracie Lou Freebush每个月都会在我们的HRIS系统中运行一份报告，并根据招募的已签署录取通知书进行检查。 她将加密文件并将其保存到我们的内部服务器。

And just like that, we have established data ownership!

就像这样，我们已经建立了数据所有权！

• Gracie Lou is responsible (or the “steward”) for compensation data.

  格雷西·卢(Gracie Lou)对补偿数据负责(或称为“管理人”)。

• The data is housed in our HRIS system and on our internal server.

  该数据被安置在我们的HRIS系统和我们内部的服务器上。

• We ensure accuracy by checking the data against offer letters.

  我们通过根据录取通知书检查数据来确保准确性。

• The data is kept secure by being encrypted.

  数据通过加密保持安全。

Easy peasy, lemon squeezy.

容易油腻，柠檬味浓郁。

有一个单一的真理来源。 或者，有一个“源代码优势等级”(我刚刚编造了但留在我身边)。 (Have a single source of truth. Or, have a “source superiority hierarchy” (I just made that up but stay with me).)

If there is a discrepancy in the data, what source will prevail?

如果数据存在差异，将以哪个来源为准？

Let’s go back to Gracie Lou in Finance. She has run her report for the month and is checking the compensation data for employees against their offer letter from recruiting.

让我们回到金融界的Gracie Lou。 她已经运行了该月的报告，并正在根据员工的录取通知书检查员工的薪酬数据。

SUDDENLY, she realizes that Victor Melling has two conflicting values for his compensation! The report from the HRIS system is showing his compensation as $180,000 while his signed offer letter is showing his compensation as $18,000.

突然，她意识到维克多·梅林(Victor Melling)的补偿有两个相互矛盾的价值观！ HRIS系统的报告显示他的报酬为18万美元，而他签署的录取通知书显示他的报酬为18,000美元。

If we have determined that our HRIS is the at the top of the “source superiority hierarchy” (because the HRIS is reviewed by the hiring manager, the operations team, the payroll team, and the legal team) we would likely conclude that there was a data entry error, aka a typo, in the offer letter.

如果我们确定我们的HRIS在“资源优势层次”的顶部(因为HRIS已由招聘经理，运营团队，薪资团队和法律团队进行审查)，我们可能会得出结论：录取通知书中出现数据输入错误(也称为错字)。

In this case, our HRIS goes through the most approvals from the most teams so we would put it at the top of our source superiority hierarchy while the offer letter has, up to this point, only been reviewed by a single person. (This example is for illustrative purposes, please, for the love of everything holy, have multiple review steps for offer letters).

在这种情况下，我们的HRIS会经过最多团队的最多批准，因此我们将其放在源代码优势等级的顶部，而到目前为止，要约信仅由一个人进行审核。 (此示例仅用于说明目的，出于对圣洁事物的热爱，请对录取通知书进行多个审核步骤)。

As your People Analytics matures and you use more and more data sources (financial data, demographic data, survey data, etc. etc. etc.) this becomes increasingly important.

随着People Analytics的成熟以及您使用越来越多的数据源(财务数据，人口统计数据，调查数据等，等等)，这变得越来越重要。

尽早建立信誉。(Establish credibility early.)

When you are first starting out, spend 80% of your time quality checking your work. 80%. Minimum. I’m not joking.

刚开始工作时，请花费80％的时间来检查工作。 80％。 最低要求我不是在开玩笑。

If people don’t trust your data, they won’t trust you.

如果人们不信任您的数据，他们将不会信任您。

Mistakes happen. Again, the data that we work with is extremely complex and it is so easy for tiny errors to run rampant in our analyses. You can be accurate 99% of the time but people will only notice the 1% when there is a mistake.

发生错误。 同样，我们使用的数据非常复杂，因此微小的错误很容易在我们的分析中泛滥成灾。 您可以在99％的时间内准确无误，但只有在出现错误时，人们才会注意到1％的准确率。

When I send a ten page quarterly report with 200 different calculations I have never received an email that says “Wow, all of these calculations are absolutely correct, you did such a good job!” However, I have received (many) emails that go something like, “the number on page 6 in row 34 should actually be 7.2 instead of 7.4.”

当我发送包含200种不同计算方式的十页季度报告时，我从未收到一封电子邮件，其中说：“哇，所有这些计算方式都是绝对正确的，您做得很好！” 但是，我收到了(很多)电子邮件，内容类似，“第34页第6页上的数字实际上应该是7.2，而不是7.4。”

I lose years off of my life every time I receive an email like this. Often, there is a good reason why the report reads 7.2 instead of 7.4. But, the horrible truth is that sometimes that person who emailed me is right, and there is a mistake. It is a horrible feeling. Chin up, Charlie. Talk to someone who can help put the mistake in perspective (i’m available for commiserating and virtual hugs) and focus on what you learned and will change going forward.

每当我收到这样的电子邮件时，我都会丧生数年。 通常，有一个很好的理由使报告读取7.2而不是7.4。 但是，可怕的事实是，有时给我发电子邮件的人是正确的，而且有一个错误。 这是一种可怕的感觉。 振作起来，查理。 与可以帮助您从错误中了解错误的人交谈(我可以用于同情和虚拟拥抱)，并专注于您所学到的知识，并将不断变化。

When there is a mistake, explain what happened for the error to occur and what stopgaps you have put in place to prevent it from happening again.

如果出现错误，请说明发生该错误的原因以及为防止该错误再次发生而采取的措施。

Here are some ways to minimize mistakes from happening as much as possible:

以下是一些尽量减少错误发生的方法：

• Compare the data to what you have calculated previously. Is this year’s turnover significantly different from last year’s turnover? Does the difference make sense?

  将数据与先前计算的进行比较。 今年的营业额与去年的营业额有很大不同吗？ 区别有意义吗？

• Have multiple people make the calculation. This can be resource intensive but can also be extremely beneficial, especially in the beginning. By doing this with a new analyst, we realized that they were calculating turnover using the average of the first and last month’s headcount while I was using the average of every month’s headcount. Both could technically be considered “correct” but we need to be consistent.

  有多人进行计算。 这可能会占用大量资源，但也会带来极大的好处，尤其是在一开始的时候。 通过与一位新分析师进行比较，我们意识到他们是使用第一个月和上个月员工人数的平均值来计算营业额的，而我使用的是每个月员工人数的平均值。 从技术上讲，两者都可以视为“正确”，但我们需要保持一致。

• Calculate by hand. Don’t hate me for this one. This is my least favorite thing to do. I love to automate and do whatever I can to make my processes more efficient. To do this, I often have templates or scripts that make calculations for me. However, I will always pick a couple of departments where I will also do the calculation by hand to ensure that the automated process is working as expected. This can be painful for those who appreciate efficiency but it is so important to ensure accuracy.

  手工计算。 不要恨我这个。 这是我最不喜欢做的事情。 我喜欢自动化，并尽我所能提高我的流程效率。 为此，我经常有为我进行计算的模板或脚本。 但是，我将始终选择几个部门，在这些部门中，我还将手工进行计算，以确保自动化过程按预期进行。 对于那些欣赏效率的人来说可能会很痛苦，但是确保准确性非常重要。

• Filter checks. When looking at high level results, it can be more difficult to spot errors. When doing your data checks, try looking at your results for smaller groups (e.g. by department, location, etc.) and ensure that their results are accurate. Errors are much easier to spot for small groups.

  筛选检查。 查看高级结果时，发现错误可能会更加困难。 在进行数据检查时，请尝试查看较小小组的结果(例如按部门，位置等)，并确保其结果准确。 对于小组来说，发现错误要容易得多。

• Create your own quality assurance checklist (example here).

  创建您自己的质量保证清单(此处为示例)。

2.使其具有可行性。(2. Make it actionable.)

There are so many amazing, wonderful, super cool things you can do in People Analytics that it can be hard to know where to begin. Start with the basics and give people something they need. To make your work actionable:

在People Analytics中，您可以做很多令人惊奇，奇妙，超酷的事情，以至于很难知道从哪里开始。 从基础开始，为人们提供他们需要的东西。 为了使您的工作可行：

• Align projects with company goals使项目与公司目标保持一致
• Have a champion有一个冠军

• Make tangible recommendations

  提出切实的建议

A great way to project plan is to start with existing business priorities and goals. In order for your hard work to have the most impact, people need to care about it and people need to be able to do something about it.

制定项目计划的好方法是从现有的业务重点和目标开始。 为了使您的辛勤工作产生最大的影响，人们需要关心它，并且人们需要能够对此有所作为。

A champion, or executive sponsor, is someone who keeps the goal at the forefront of people’s attention, holds people accountable, and promotes the work being done. Ideally, this person is highly visible within your organization and has some “skin” in the game (i.e. they are personally or professionally motivated to see the goal accomplished).

拥护者或执行赞助者是指将目标放在人们关注的最前沿，让人们负责并促进完成的工作的人。 理想情况下，此人在您的组织中是高度可见的，并且在游戏中具有某些“皮肤”(即，他们是出于个人目的或专业动机来实现目标)。

Get this person bought into your project as early as possible. This should be a mutually beneficial relationship: you help the sponsor achieve the goal and the sponsor helps promote your work. Without an executive sponsor, a project can quickly die out with no one using, reading, or caring about your work. Sad. Day. A champion can make a world of difference.

尽早让这个人加入您的项目。 这应该是互惠互利的关系：您帮助赞助商实现目标，而赞助商则帮助促进您的工作。 没有执行者赞助，一个项目可能很快就会消失，而无需任何人使用，阅读或关心您的工作。 伤心。 天。 冠军可以改变世界。

Finally, you need to make tangible recommendations based on the results of your project.

最后，您需要根据项目的结果提出切实的建议。

When I first started, I would present my findings and then expect people to go forth and prosper. I thought my job was to run the analysis and their job was to run with it. I would send out my findings expecting the world to change. But, to my chagrin, nothing happened.

当我刚开始的时候，我会提出自己的发现，然后期望人们前进并繁荣。 我以为我的工作是进行分析，而他们的工作是进行分析。 我会发表我的发现，期望世界会改变。 但是，令我恼火的是，什么都没有发生。

There was a crucial gap — I was sharing my findings but people still didn’t know what to do with them. Okay, we found that education level was not a significant indicator of performance for a given position, so what?

有一个关键的差距-我正在分享我的发现，但是人们仍然不知道该如何处理。 好的，我们发现教育程度并不是给定职位表现的重要指标，那又如何呢？

We need to help our audience understand, not only our findings, but what our findings mean, and what they can do about it.

我们不仅要帮助听众理解我们的发现，还需要帮助听众理解我们的发现意味着什么，以及他们可以做什么。

If we found that education level was not a significant indicator of performance, this might mean that we should change our selection process for that position. By widening our selection criteria, we could increase our number of candidates by X%. This would also mean we could save $Y by being able to hire someone with a lower education level, who would likely have a lower compensation, but be just as effective on the job.

如果我们发现教育程度不是绩效的重要指标，则可能意味着我们应该更改该职位的甄选过程。 通过扩大选择标准，我们可以将候选人人数增加X％。 这也意味着我们可以通过雇用较低文化程度的人来节省$ Y，这些人的薪水可能较低，但在工作中同样有效。

Now they are picking up what you’re putting down. They’re buying what you’re selling. Smelling what you’re cooking… you get it.

现在他们正在接您要放的东西。 他们在买你在卖的东西。 闻到您正在做的菜...明白了。

3.使它易于访问(理想情况下，也使其美观)。 (3. Make it accessible (ideally, make it pretty, too).)

Before you panic, you don’t have to go out and learn a data visualization software (at least not yet…). This can be simple. It can be in Excel (there is a reason why everyone and their mom uses it, there is no shame in that game).

在您惊慌之前，您不必出门学习数据可视化软件(至少现在还没有…)。 这很简单。 它可以在Excel中使用(每个人和他们的妈妈都使用它是有原因的，在该游戏中没有羞耻感)。

Just make it visual, and ideally, make it interactive. Humans have superpowers when it comes to digesting visual information. We do it extremely quickly and efficiently. Take advantage of our natural gifts!

只需使其可视化即可，理想情况下，使其具有交互性。 在消化视觉信息方面，人类拥有超能力。 我们非常快速有效地做到这一点。 利用我们的天然礼物！

When sharing your data make it:

共享数据时，请进行以下操作：

• Visual视觉效果
• Interactive (e.g. with filters)互动式(例如带有过滤器)
• Readily accessible (e.g. real-time data access)易于访问(例如实时数据访问)

Let’s try out the two experiences below and think about which one our audience would prefer:

让我们尝试下面的两种体验，并考虑一下我们的观众更喜欢哪种体验：

选项1： (Option 1:)

There is a lot of information here, which could be a good thing. If your audience is someone who is immersed in this data all the time, they might need this level of detail. But it also difficult to pull meaningful takeaways from this data. There is a lot here, but what am I looking for? How much turnover is “bad”?

这里有很多信息，这可能是一件好事。 如果您的听众是一直沉浸在这些数据中的人，那么他们可能需要这种详细程度。 但是，也很难从这些数据中获取有意义的收获。 这里有很多东西，但是我在找什么呢？ “不良”营业额是多少？

选项2： (Option 2:)

Option 2 is a more visual option.

选项2是更直观的选项。

It is extremely simple, but it is already a huge improvement over where we were before. Now we can filter to the department we are interested in and our bars change color based on whether they are above or below the average monthly turnover for that department.

这非常简单，但是已经比以前有了很大的改进。 现在，我们可以过滤到感兴趣的部门，并且条形图会根据它们是否高于或低于该部门的平均每月营业额来更改颜色。

We can see immediately that in January and April the Marketing team had higher turnover than they usually do. April turnover doesn’t look much higher than usual, but January does and might be worth some more investigation.

我们可以立即看到，在1月和4月，营销团队的营业额比平时更高。 四月成交看起来并不比平常高很多，但确实月份，可能是值得一些调查。

For example, we might find that the majority of these people left right after bonuses were paid out. It might then be worth pushing back when bonuses are paid out in order to help our retention.

例如，我们可能发现这些人中的大多数人在奖金发放后就离开了。 因此，在发放奖金以帮助我们留住人才时可能值得推迟。

Things are really getting exciting now.

现在事情真令人兴奋。

互动性 (Interactivity)

Leaders tend to only care about the data that is “relevant” to them. If you send a leader full company turnover, they will say okay, fine, but what is the turnover for *my* department?

领导者往往只在乎与他们“相关”的数据。 如果您向领导者发送公司的全部营业额，他们会说很好，但是，我的部门的营业额是多少？

When sharing data it is important to make your results interactive enough so that leaders feel they can narrow in on a group that they feel is specific enough to them.

当共享数据它使你的结果互动足以让领导觉得他们可以缩小在一组，他们的感觉是不够具体，对他们来说是很重要的。

Every leader is their own special snowflake. Some like to be able to filter by department or cost center, some by location, and some even by different positions.

每个领导者都是他们自己的雪花。 有些人喜欢能够按部门或成本中心进行过滤，有些人则可以按位置进行过滤，有些甚至可以按不同的职位进行过滤。

However, it is our job to save leaders from themselves and not make our results so interactive that they are no longer meaningful.

但是，我们的工作是将领导者从自己手中救出来，而不是让我们的结果如此互动以至于它们不再有意义。

For one, it is critically important to keep things like confidentiality in mind (e.g. not showing results for groups with less than 20 people). Results also just lose their meaning when the size gets too small. 100% turnover would feel really alarming, but not if there was only one person in that department/location/position.

首先，牢记诸如保密之类的事情至关重要(例如，对于少于20人的小组，不要显示结果)。 如果大小过小，结果也将失去意义。 100％的营业额确实令人感到震惊，但如果该部门/位置/职位只有一个人，则不会。

Help educate leaders on why you are limiting the amount of interactivity in your results and then be firm with the decision.

帮助教育领导者了解为什么要限制结果中的交互性，然后坚定决策。

易于访问。 (Readily accessible.)

You could have the coolest, most groundbreaking analysis ever completed. But if no one sees it — if no one has access to it — it doesn’t count for sh*t. People can’t use what they don’t have.

您可能会完成有史以来最酷，最突破性的分析。 但是，如果没有人看到它-如果没有人可以使用它-它就不算什么。 人们无法使用他们没有的东西。

Making data readily accessible is the idea of data democratization. This doesn’t mean that every employee should have full access to all raw data that goes into your analytics.

使数据易于访问是数据民主化的想法。 这并不意味着每个员工都应该拥有对您分析中所有原始数据的完全访问权限。

This means getting information in the hands of people who need it at a level of access that is appropriate.

这意味着需要以适当的访问级别将信息掌握在需要它的人的手中。

Good things happen when more people have more access to data… for the most part. Obviously always consider the privacy considerations, ethics, and potential consequences of giving access to data (I’m talking to you, Zuckerberg).

当更多的人可以更多地访问数据时，好事就会发生。 显然，始终要考虑允许访问数据的隐私权，道德和潜在后果(我正在与您交谈，扎克伯格)。

When building a People Analytics function, it is critically important to decide how you are going to share your results. Important considerations are:

构建人员分析功能时，决定如何共享结果至关重要。 重要的注意事项是：

• Content (e.g. dashboard, report, etc.)

  内容(例如仪表板，报告等)

• Format (e.g. Excel, BI tool, HRIS system, etc.)

  格式(例如Excel，BI工具，HRIS系统等)

• Cadence (e.g. monthly, quarterly, etc.)

  节奏(例如每月，每季度等)

Regardless of the decisions you make, a leader should know exactly how, where, and when they can access their results. These decisions can, of course, be iterative. In the beginning you might be sending out reports in Excel monthly and then, as your People Analytics function matures, you might give leaders real-time access to dashboards in Power BI.

无论您做出什么决定，领导者都应该确切知道如何，在何处以及何时获得结果。 这些决定当然可以是迭代的。 开始时，您可能会每月在Excel中发送报表，然后，随着People Analytics功能的成熟，您可以为领导者提供对Power BI中仪表板的实时访问。

Whatever you do, make your data a breeze to digest. Heck, make it a JOY. It will keep people coming back for more.

无论您做什么，都使数据轻而易举地消化。 哎呀，做个快乐。 它将使人们回来更多。

In summary, when building a People Analytics function, start by doing the following:

总之，在构建人员分析功能时，请先执行以下操作：

1. Make it accurate
   - Establish data ownership

   使其准确-建立数据所有权

   Make it accurate
   - Establish data ownership- Have a single source of truth

   使其准确-建立数据所有权-拥有单一的事实来源

   Make it accurate
   - Establish data ownership- Have a single source of truth- Establish credibility early

   使其准确-建立数据所有权-拥有单一的事实来源-尽早建立信誉

2. Make it actionable
   - Align projects with company goals

   使其可行-使项目与公司目标保持一致

   Make it actionable
   - Align projects with company goals- Have a champion

   使其可行-使项目与公司目标保持一致-有冠军

   Make it actionable
   - Align projects with company goals- Have a champion- Make tangible recommendations

   使其可行-使项目与公司目标保持一致-拥有冠军-提出切实建议

3. Make it accessible
   - Use visuals

   使其可访问-使用视觉效果

   Make it accessible
   - Use visuals- Make it interactive

   使其可访问-使用视觉效果-使其互动

   Make it accessible
   - Use visuals- Make it interactive- Make it readily accessible

   使其可访问-使用视觉效果-使其具有互动性-使其易于访问

Keep banging your head against the wall. It gets so much better!

继续用力撞墙。 它变得更好了！