There’s been a lot of talk about “big data” recently. The idea that if we just collect more and more data we can find hidden correlations and exploit those for our advantage. It’s definitely an interesting field and will shape many larger organizations but you don’t need “big data” to succeed.
I am a proponent of what I call data informed decision-making. Too often we rely on instinct, our gut, or randomness when making strategic decisions. It’s not that these things are always wrong but I believe the organizations that truly succeed don’t rely completely on instinct, they move towards data informed decision-making. Integrating data informed decision-making into your organization is quite simple. The first step is to simply begin collecting data that you feel is informative in some way. This includes things like sales & profits by region and data but can move much beyond that. Especially in the social sector it is important to move from measuring activity data to outcome data.
Next, when making key strategic decisions, go back to the data. Ask yourselves can we justify this decision based upon the data in front of us? If not, can we collect some data to help support this decision? Data informed decision-making is just about adding a check in the decision-making process. It’s about checking your instinct, gut, or ideas with data. You won’t always be able to support every decision with data but this will help your organization succeed more of the time.
Stanford Social Innovation Review highlighted the work of Data Without Borders (DWB) last week. The article talks about how social organizations play an important role in the data collection universe but rarely have the resources to house data analysis talent internally. Data analysis is hugely important when it comes to operations and impact which is why I started Means Well Does Good in the first place. I love getting to work with organizations and operate as an outsourced data analysis department for them.
Be sure to check out the article here.
I have been spending a lot of time working on logic models lately. It’s a tool that I believe every nonprofit must utilize. Logic models are quite simple but their effect can be quite powerful because they help to connect activities to outcomes. They help ensure that all the work you are doing is connected to your mission and being done for some impactful reason.
The best way to begin your logic model is with a problem or opportunity statement. I like to think of this as how the current reality. For example it could be that X% of kids in a certain community are homeless or hungry. You then write the Long-Term Impact you want to have on that problem. This is how the world should look after your organization is successful. So for example, there are no more hungry kids.
Then you think back a step and say, for that long-term impact to occur, what kinds of changes in behavior, knowledge, or circumstance need to change for that to occur. It could be things like everyone who is eligible for food-stamps is on them. These are your near-term impacts.
Backing up one step further what kinds of outputs need to occur for that change to take place. This is often a specific, benchmarked number like X number of people need to become educated on food stamps. Backing up one step further, what activities need to happen for that output to take place. There need to be people to teach people about food stamps and help them get registered. The final step is examining the inputs necessary to perform those activities. That includes financial resources, human capital, facilities, etc.
So in the end you have created a causal change that looks like this.
There is a problem in the world.
If we can gather these inputs we can perform these activities.
If we perform these activities we will achieve this level of output.
If we achieve that level of output we will realize these near-term impacts.
If we can realize these near-term impacts we will have this long-term impact.
In this way everything that your organization is doing is leading to some level of impact. This is a great way to focus your resources and energies around things that will actually be making a difference.
A great post by The Economist discusses the idea that randomized control trials in economics might not be the gold standard some believe them to be. They specifically address the outcome of a trial by an agricultural economist who did two studies on an improved form of seed. Test 1 was a double blind randomized control trial in which the participants did not know whether they were in the test or control group (i.e. they did not know whether they received the new seed or not). There was no difference in yields between the two groups. In test 2 the participants were aware that they were receiving the new high yield seed and it significantly beat the control group in terms of yield.
This is quite an interesting result. On the one hand it makes sense that as an input in your production function changes (the seed) you should adjust other inputs like water, land quality, etc. But the starkness in the result is startling. The article argues that many of the benefits seen in the tests like those of #2 do not really get at the true outcome.
For more read the article here.
It’s no secret that I’m a fan of data. I really feel that it can be used to help in a wide range of sectors. A recent Atlantic Cities post entitled How to Catch a Criminal with Data talks about an initiative between the University of Memphis and the city’s police department to use data to solve and predict crime. This idea has fascinated me since I first starting watching Numb3rs, years ago. During grad school I was always interested in applications of math and game theory but what makes the Memphis initiative so interesting is that it is much simpler than all of that.
The biggest thing they’ve done in Memphis is make the data they already had searchable.
For example, the article talks about a child abduction that was witnessed by two 10 year olds, not always the most reliable witnesses. On as little information as a car color, number of doors, and vague personal description the police department identified a possible suspect and the girl was recovered 30 minutes after her abduction. Thirty minutes! And you didn’t need someone with a PhD, it was all completely reliant upon existing data that the Memphis police department always had but was never in useful form.
As I delve further into the data world this issue of how to best organize data keeps coming to the surface. It is such a huge and interesting problem and one that is only getting easier to solve over time. It will be exciting to see all the possibilities that unfold in the future.
The hall fills. GM’s from some of sports greatest teams take their place on stage. Crews from ESPN are recording, podcasting, and live-blogging the event. No this is not the ESPY’s or a draft, it’s a sports statistics conference at MIT’s Sloan School of Management. Up to 2,200 attenders from 1,500 last year, the conference, in it’s 6th year, attracts the nerds of the sports world. Guys and girls who sit behind computers cranking out new statistics and methodologies to understand and predict athletic performance.
The conference draws guys like Kirk Goldsberry who has a PhD in geography and usually publishes on topics in public health but who applied his geographic statistic techniques to understand scoring percentage in the NBA based on court position. He woke up one day never having been on network TV, in the New York Times or Sports Illustrated, and went to bed the next night having accomplished all those and more.
Sports analytics is something I’d definitely like to learn more about so if any readers have suggestions on where to begin I’d greatly appreciate it. Check out Fast Company‘s article on this years conference entitled In Relentless Jock-Nerds War, Hope for Peace Through Analytics for more info.
I think that more organizations need to celebrate. They need to bring out the party balloons, turn up the cheese factor, and make some silly trophy’s Dundies style.
Yesterday I had the opportunity to engage in one such celebration. It was definitely cheesy. It was definitely silly. But I think it was one of the greatest things that the leadership did to shape their organization. Celebrations are great at showing what you really care about, where you’re really going, and motivating employees to go the extra mile.
For example, the organization I had the opportunity to celebrate with is hyper-focused on financial sustainability right now. The award recipients all had made progress on that front and that was highlighted. It wasn’t the only thing that was highlighted but it was a resounding theme.
Celebrations also help you mark progress. Too many organizations get stuck in the day to day operations that they forget they are moving in any direction whatsoever. Stopping to celebrate progress is a great reminder of how far you’ve come while also setting up the vision for how far you have to go.
So stop, make some trophy’s, come up with some silly rewards, and celebrate!
Restricted giving should be outlawed. I know, BOLD, right? But come on, since when do we think the professionals don’t know how to allocate resources? If you can’t even trust an organization to do that, should you really be giving them money? Restricted giving is essentially a selfish act, saying that you the donor knows more about the proper allocation of resources than those running the organization.
If you are unfamiliar with the concept, restricted giving is when the donor specifies how their donation should be spent. For example they might specify that it go towards a certain program or cost. We would never do this in the for-profit world, why must non-profits deal with this? Could you imagine telling a company that you would buy their product as long as the income was allocated to production but not to marketing because you felt that was wasteful.
What are your thoughts? Do you hate restricted giving as much as me? If not, why not?
Inherently, data is a collection of the past. It is a summation of events that have already happened. We use data to help predict the future but data is inherently about what has come before. Can data then, be innovative?
I began thinking about this question after reading a post from Fast Company Design entitled, To Innovate, You Have to Stop Being a Slave to Data. The article talked about how companies are more and more using customer data as they develop new products and since people are inherently averse to change you can’t actually create anything new. I believe it was Steve Jobs who was the mindset that you shouldn’t ask customers what they want when developing a new product but should create something that they don’t realize they want. I think that is a good idea to think about when innovating.
So how can data play a part in innovation? First, I think data helps organizations understand their current reality. They can see what is currently resonating with consumers and what is not. Second, data can help predict what tomorrow might look like, thus helping innovators know what might be around the next corner.
At the end of the day though, I think data cannot produce innovation. I think it can play a huge role in the process of innovation but it is not at the steering wheel creating the next great product, idea, or organization. It cannot be ignored but you must also understand that it is fundamentally a relic of past events which is not always a good indicator of the future.
Recent Comments