Process vs. Outcome (People Analytics)

Loading...

來自 University of Pennsylvania 的課程

商业分析毕业项目

318 個評分

University of Pennsylvania

商业分析毕业项目

318 個評分

課程 5（共 5 門，Specialization 商业分析）

The Business Analytics Capstone Project gives you the opportunity to apply what you've learned about how to make data-driven decisions to a real business challenge faced by global technology companies like Yahoo, Google, and Facebook. At the end of this Capstone, you'll be able to ask the right questions of the data, and know how to use data effectively to address business challenges of your own. You’ll understand how cutting-edge businesses use data to optimize marketing, maximize revenue, make operations efficient, and make hiring and management decisions so that you can apply these strategies to your own company or business. Designed with Yahoo to give you invaluable experience in evaluating and creating data-driven decisions, the Business Analytics Capstone Project provides the chance for you to devise a plan of action for optimizing data itself to provide key insights and analysis, and to describe the interaction between key financial and non-financial indicators. Once you complete your analysis, you'll be better prepared to make better data-driven business decisions of your own.

從本節課中

Module 3: Your Strategy

In Module 3, you will focus on creating your recommended strategy for GYF to address adblockers. Your strategy does not have to be lengthy, but it must be clear, and it must address the problem. (Hint: if you have a clearly defined problem, your strategy is much more likely to be clearly defined as well). You'll be submitting your strategy for peer review, and then also reviewing the work of at least 3 of your peers. It's OK if reviewing the strategies of other learners in this course gives you further ideas for revising your own strategy. One of the primary benefits of peer review is to expand the range of feedback you can get, and we designed this Module around peer review so that you can get as much feedback as possible before moving on to the next phase of the project. You may find the resources and lectures below helpful in formulating your strategy and considering how data can be leveraged and appropriately understood. You are strongly encouraged to complete and include your response to Application Exercise 2 as part of your Strategy.

Performance Evaluation: the Challenge of Noisy Data (People Analytics)6:46

Finding Persistence: Regression to the Mean (People Analytics)11:56

Extrapolating from Small Samples (People Analytics)5:53

The Wisdom of Crowds: Signal Independence (People Analytics)5:56

Process vs. Outcome (People Analytics)7:47

Hiring 1 (People Analytics)8:50

Hiring 2 (People Analytics)9:07

與講師見面

Wharton Teaching Staff
Educators
The Wharton School

So the fourth and final issue we're going to talk about

on performance evaluation is process versus outcome.

And the first issue that we want to add here

is that a firm ought to, and individuals ought to consider

a broader set of objectives than they typically consider.

So organizations usually care not just about what exactly happens, but

how it happens.

How a person goes about their job.

So for example, perhaps most importantly, along the way to creating some outcome,

what impact did that employee have on other employees?

What impact did they have on others?

Can you fold that into the performance evaluation as well?

This just highlights that we

care about more things than just what is traditionally measured, and

so one prescription that comes from it is to measure more things.

An example of why we do this is research by Bond, Carlson and

Keeney, they found that people consider too few objectives.

And this is in negotiation settings, this is in decision making settings, that

if you ask people what they're trying to accomplish, they'll list a few objectives.

But then if you ask others, you ask a large group and

then you share all the possible answers with people, people will go back and

say, yeah, I forgot that, and I'll add that.

And in the end, they list it on their own,

something about half of all the objectives they ultimately recognize as relevant.

So, left to our own devices, we're a little too casual,

a little too informal about focusing on a few narrow issues instead of a broad set.

Now, it can lead firms to focus too narrowly.

An example of a firm that recognized this, Dell Computers in the early 2000s,

famously hard-charging, famously results-oriented,

changed their performance evaluations.

They changed from 100% results based to 50% results based,

what an employee accomplished, and 50% how he or she accomplished it.

So they're judging not only what the person does, but

their impact on other people.

Because they had too much experience with managers who were running over people

to hit the numbers to get the bonus, and in fact

the firm cared about things other than just the numbers that were being measured.

In general, the more uncertainty there is in an environment,

the less control an employee has over the exact outcomes.

The more a firm should emphasize process over outcome.

So the more noise, the less control the employee has,

the more they should be evaluated on process, and not on outcome.

One way you can go about this is to use analytics to figure out

which processes tend to produce the desired outcomes.

And what you're looking for

here is what's the more fundamental driver of the outcome.

It could be that you're measuring only the last step in the process.

And in fact, there are some important intermediate steps

that are more fundamental and they might provide additional performance measures.

So, for example, the sports analytics world gives us an illustration of this.

In hockey, for a long time, teams were evaluated, and,

in fact, players were evaluated based on goals.

And if you were trying to figure out whether your team was really good or poor,

they would look at the number of goals they'd been scoring.

If they wanted to evaluate whether a player was strong or weak,

they would look at his contribution to the goal scoreboard while he was on ice.

And this is fine and it's related and it's important, but can you do better?

It turns out there aren't many goals scored in hockey.

And sometimes they are scored because they hit the pole and

went left when others aren't scored because they hit the pole and went right.

They tip off of people's skates, all kinds of crazy things happen.

There's a lot of noise between what a player controls and what actually happens.

And whenever there's that noise,

you want to be careful about how much weight you put on it.

It's not giving a very reliable signal for the true effort or

the true ability underneath it.

So what do they do?

They determine through analytics that a better predictor was not goals but shots.

That if you looked at how many shots, they call it shots on goal, shots that were hit

basically at the goal, at the net, that that was a better predictor.

That was a more reliable measure, and

one would think about it is, it's more persistent.

That a guy, or a player, or a team that looks good

in one period on shots is more likely to look good on next period on shots

than a player who looks good on goals is to look at on goals in the next period.

It's a more persistent and fundamental measure.

Well, you can take that even further, they subsequently realized

what really matters is possession, it's not even shots.

So there's, again, so much noise on shots that,

the best measure they can find on team or player is contribution to possession.

The teams that keep the puck,

the teams that had the puck are the ones that are most likely to shoot.

The ones who are more likely to shoot are most likely to score goals.

But, they had to figure that out going backwards, looking for the more and

more fundamental measure, and you can think of these as process measures.

They're getting away from the noisy, rare outcome measure of a goal

to the more fundamental, more reliable measure, process measure, possession.

What about non hockey applications you might reasonably ask?

Well, it's not hard to come up with them.

One, I'd push you to think about what that means for your organization,

but one conversation I've had along these lines is with sales organizations.

Can they come up with more fundamental measures than the traditional dollars

booked?

Can you, for example, consider the process that leads to that dollars booked?

is it the number of bids that a salesperson gets her organization into or

maybe before that, is it the relationship building that

the sales person does in order to get the bids in order to book the dollars?

Or maybe even it's earlier than that, it's an earlier process yet.

It's the number of contacts generated, and we don't know exactly

which of these would be most persistent, it will vary of course by organization and

by industry, but the idea is to take data to the problem.

And determine where in the process you might start adding assessments

to get away from these relatively rare noisy outcome measures.

At the very least,

to supplement them with more fundamental drivers earlier in the process.

I want to end this section with a quote from Shane Battier,

another sports example, another Michael Lewis example.

This was an article written in the New York Times Magazine

a few years ago, where Shane Battier was playing for the Houston Rockets and

the Rockets were famously users of data analytics.

And Battier was famously a player who, despite not having any obvious

flashy stats or obvious skills, hung around the NBA.

Teams found him valuable, and

he became even more valuable when he went to the Rockets.

They started looking at process.

They started coaching him on,

here's something you can do differently that's going to influence the game.

If you defend a player for

example this way, you're going to decrease the likelihood that he makes the shot.

If you play offense this way,

if you attack the post in this way, you're more likely to score.

So it's always likelihood with the Rockets.

It's always likelihood in sports analytics, and it allowed Battier,

it kind of freed him up to focus on what he knew was right, and not worry so much

about whether it happened to turn out well on any particular trip down the court.

Or happened to turn out badly on another trip down the court.

He could focus on the process.

He put it this way.

Knowing the odds,

Battier can pursue an inherently uncertain strategy with total certainty.

He can devote himself to a process and

disregard the outcome of any given encounter.

This is critical because in basketball, as in everything else, luck plays a role, and

Battier cannot afford to let it distract him.

探索我們的目錄

免費加入並獲得個性化推薦、更新和優惠。

Coursera 致力於普及全世界最好的教育，它與全球一流大學和機構合作提供在線課程。

© 2018 Coursera Inc. 保留所有權利。

通過 App Store 下載

通過 Google Play 獲取