This year Futurice experiments with charity projects with a budget of half a million euros as an extension to our open source and social impact program. The investment is mainly made as work by our employees. In January we asked for project applications through a simple web form. We received about 30 of them. Some of them are quite good! This article describes our selection procedure so far, the current state of affairs, and what will happen next.
You can read this blog post for a lengthy description, or download this free ebook for an even lengthier one, but to sum it up: Because our people want us to make the world a better place and doing so creates a lot of value to the company.
All our social impact activities are based on a simple value trifecta. Our actions create value to the acting individuals, the company, and the society. All three. Our grading categories are there to address this.
In addition we want this to be a program for the people and by the people. In a non-tiny company you almost always deal with a conflict between how to involve as many people as possible, and how to make decisions efficiently. In our case the efficiency can suffer, as it means we give our people real opportunities to actively participate.
Also we are currently working on three charity projects, selected last year as pilots for this initiative. So we can take our time with this selection round.
During February we graded applications. Our grading categories are:
Social impact – is this a problem worth solving?
Competence development – what professionally relevant learning would there be?
Student cooperation potential – can we do this with a group of students?
Partner cooperation potential – would this interest our customers?
Open source potential – will it be open source and how valuable would it then be to others?
PR potential – does it make for a good story?
These categories mirror the value we want to create. Social impact and open source are value to the society and to some extent also to the individuals. Competence development is value to the individuals and the company. PR, student, and partner opportunities are also good company value.
Each category is estimated with a Fibonacci sequence compatible grade: 1, 2, 3, 5, and 8 being the highest. In addition we estimate:
Since we go through a lot of applications we have to compromise on how thorough our first assessment is. We do not calculate person days of work at this point, nor do we carry out any proper risk analysis. That happens later, if the project is selected to a feasibility study.
Then to get the grade we unleash math!
((Social Impact * Multiplier) + Competence Development + Open Source Significance + PR Opportunities + Student Cooperation + Partner Cooperation) - (Effort * Risk)
We sum up the positive grades, giving some extra weight to social impact. Then we deduct effort multiplied by risk of the score. The exact formula used doesn't really matter so much. The grading is relative and what matters is that it treats them fairly, as we want to establish an order for the applications.
So we sum up the good things and deduct effort and risk to grade a project.
First the facilitators for this process, Sebi and I, each graded all the categories for all the applications. Then we averaged our scores.
Where our grades varied too much, a discussion a bit like this may have taken place:
Sebi: So you think helping endangered tree kangaroos by hybrid bioacoustic 4D quantum printing has an extremely low social impact.
Sebi: Care to elaborate why?
Teemu: Wellllll you know I really don’t like kangaroos, when I was a child there was this horrible cartoon Kössi Kenguru-
Sebi: Dude. What? Also tree kangaroos are just the initial target. This can be applied to all endangered fauna! Even Mediterranean monk seals.
Teemu: Monk seals?? My totem animal!!! <3 <3
Sebi: Yes, as we all know. So?
Teemu: Okay, I’ll raise my social impact grade from 1 to 5.
The resulting grades allowed us to pick the top twelve projects to the proper grading process. This filtering was necessary, because the grading process is intentionally quite heavy to allow for wide organisational participation.
We assembled temporary teams of four to six people, to grade a selected category for each of the twelve projects. The teams used the same fibonacci grading scale. The selected approach? Planning poker.
A somewhat imaginary Open Source Impact grading scenario with planning poker:
Teemu: Okay, you have spent ten minutes reading about this project and I have answered some questions. Now, the open source impact. What do you expect the potential to be?
Ville: … tree kangaroos?
Teemu: Yes, yes! But nevermind that. Think monk seals. Let’s go, let’s go! Cards on the table!
Daniel reveals a card with the number 8.
Ville reveals a card with the number 3.
Leena reveals a card with the number 5.
Matias has nodded off.
Matias reveals a card with the number 2.
Teemu reveals a card with the number 5.
Teemu: Okay, quite some difference between the lowest and the highest. Daniel, please explain yourself?
Daniel: This right here is shit of the dopest kind. Nobody has ever even heard of hybrid bioacoustic 4D quantum printing before and we are going to open source it? It’s going to be bigger than BeOS.
Teemu: Any other thoughts? No? Okay! Revote!
Daniel reveals a card with the number 8.
Ville reveals a card with the number 5.
Leena reveals a card with the number 5.
Matias reveals a card with the number 5.
Teemu reveals a card with the number 8.
Teemu: Close enough to count. Let’s move on to the next one!
This was done for all the categories. Social impact, competence development, open source impact, student cooperation, partner cooperation, PR potential, effort, and risk. We graded social impact and effort & risk three times with different teams at different sites to get a wider perspective on the most important categories.
Let’s look at some real data. Our actual project grades:
We decided to take the top six forward to the final round of the selection process, the employee endorsement.
Here are those top six projects described with a single paragraph each. They are not in the order by their grades.
People who have as a child been customers to the child services, such as foster care, often lack the social support networks that people rely on in their own parenthood. There is often also an identifiable distrust towards authorities. These factors result in a higher probability for their children, in turn, to require assistance from the child services and protection. A service that helps people to find peer support and ask experts some questions anonymously would help in providing the missing support.
International arms trade is not very transparent. Knowledge about how the arms move allows the creation of better arms control mechanisms. In Finland this information is annually put together from different sources by a group of researchers. The impact of their work would greatly increase, if this complex data could be presented to people through modern digital means.
People often discuss prescription drugs online. How closely they follow the doctor's instructions, what they think about the drugs, what kind of advice they give to each other… how they relate to the medicine, when the doctor is not present. A large database of Finnish forum discussions exists and can be used for data analysis to refine valuable information regarding this subject to be shared with researchers, doctors, and other relevant parties.
The presence of experts in the media has increased notably during the last two decades. The problem is that the diversity of the experts commenting in the media is really low. Even though there is a large reserve of expertise in Finland, the experts utilized are often the ones that already appear frequently in the media. This carries both quality and credibility risks for the media. It is also difficult for many experts to get their voice heard in the public discussion. This is a match making problem that can be at least partially resolved with modern technology, allowing the journalists and experts to find each other.
Baltic Sea is the most polluted sea in the world. Governments alone do not work fast enough to remedy the situation. There is a lot of public interest towards protection projects, but a suitable crowdfunding platform that allows to curate and manage this process is missing.
The scout education activates individuals in relation to themselves, other people, the society, and the environment. To support this, a lot of digitised information about activities and goals exists. A mobile app would help to better utilise this information in the modern world. It would also allow the collection real-time usage information, allowing the scout movement to improve and modernise their education.
The final order of these projects will be determined by company wide employee endorsement. A web form will contain the relevant information for each of the six projects, and our people can decide whether to endorse them.
The earlier grade calculated for the projects no longer defines their order. It is just information given to the people who endorse these projects and it is up to them individually what impact it has. We are confident enough that these top six projects are all problems worth solving.
However, we cannot take six charity projects on right from the start. This is why we need to establish their relative order.
So we’ll start with the one that gets the most endorsement votes by our people?
Well, perhaps. It is not quite that simple.
What still matters from the earlier grading is the effort and risk metrics. We want our social impact to be as high as possible. We also have limited resources. If the highest endorsed project is huge in effort and risk, maybe even bigger than the rest of them together, might it make sense to do the less endorsed projects instead, to maximise the impact?
An imaginary example:
We have 65 effort points to spend. Which projects should we choose, to optimise the impact?
Yeah I didn't bother to work it out either. In reality it’s even more complex than this. Different projects require different types of competence, such as machine learning, 3D design, basic web programming, native Android... We have the whole skill versus availability thing to consider.
Also we may be able to draw on some external resources, helping with some of the projects. Customers may also have differing schedules that are easier or more difficult for us to accommodate.
To summarise; employee endorsement carries a lot of weight, but the resulting order is not followed blindly.
We'll let the endorsement poll run for a few days. Then we decide the order. For each project that goes forward this means a few weeks of intensive analysis of conceptual and technical risks. Discussion of the cooperation model, understanding customer expectations, early backlog planning, schedule and resource estimation, and so forth.
After this feasibility study there’s a go / no-go decision for each project. Then we move on to the design and implementation, according to the schedule mutually agreed with the customer.