Open Science Adoption within Organizations: Lessons for individuals from theories about technology adoption
Eli Holmes
NOAA Fisheries
About me:
NOAA Fisheries (NMFS) Science Centers
NWFSC
AFSC
SWFSC
SEFSC
NEFSC
PIFSC
NMFS Openscapes Mentors Program
2021-2022 Four NMFS Openscapes Champions Cohorts
2023 NMFS Openscapes Mentors Program
https://github.com/nmfs-openscapes
GRASSROOTS EFFORT
“The Open Science/Open Source train has left the station. Get on board or get left behind!”
– Summarizing the participant chats during a public NASA presentation on their Open Science initiative. There was a train in the background at one point during one of the presentations.
How do innovations spread through organizations? How can Diffusion of Innovations theory help us with the next steps?
EM Rogers (1962) “Diffusion of Innovation” theory
Predictable progression of stages as idea diffuses through a population: innovators, early adopters, early majority, late majority, laggards.
Time
Each group has different personalities, different motivations, and different objectives.
Tipping Point
There is a tipping point around 15% when the idea is widely spread enough that it begins being adopted faster. Between 3 to 15% adoption, progress is slow and process goes through a painful “Trough of Disillusionment”.
Time
Tipping point
The Early Adopters are critical
Time
Early Adopters: Characteristics
Attributes of OS Early Adopters
Like new technology, but are a bit more practical than the ‘Innovators’
Embrace change and comfortable with new ideas
“How to” manual is enough for them. Don’t need convincing.
See the potential of the new idea even in the fairly early stages and have the temperament (doggedness) to develop ideas into full working implementations
Challenges for OS Early Adopters
Isolated. Each center only has a few ‘Early Adopters’
Differential support for OS at each center.
Not in leadership positions, though may be team/project leads. Many tend to be junior.
Embedded in teams that are not using OS workflow
In my agency, we are in the early Early Adopters stage of adopting Open Science (OS)
OS Early Adopter stage
OS Innovator stage
There are a number of common Open Science & Open Source languages. Python & R are very common. R is more common in fisheries while Python more common in the earth sciences and data science fields (my impression). Both have large and active Open Science & Open Source communities.
Identify, support and develop the Early Adopters
Their energy and effort is what drives the initial diffusion process
Q1. Where should we put our energy right now – in prep for the mentor program?
Pre-2018 or so, very hard/impossible for OS Innovators to collaborate
Literally no way to get information to each other
NWFSC
AFSC
SWFSC
SEFSC
NEFSC
PIFSC
The Early Adopters are hard to reach and many are isolated
Early Adopter
Around 2018-2019, FIT begins filling the gap for providing R and software dev training. NMFS-wide email list. Interestingly, Google email integration also made it much easier to communicate across NMFS.
NWFSC
SWFSC
SEFSC
HQ
NEFSC
PIFSC
AFSC
2021: R Workflow Workshops opened up to all NMFS staff and advertised using FIT email list. Openscapes NMFS cohorts in 2021 involved 4 science centers. NASA Openscapes GitHub org (early 2021) provides a framework for moving forward. NMFS R UG (Nov 2021) joined by 100+ NMFS scientists across all centers
NWFSC
SWFSC
SEFSC
HQ
NEFSC
PIFSC
AFSC
NMFS R UG has members across all science centers. No more silo-ing!
2022 Main Goal: Reach the Early Adopters
OS Early Adopter does not mean they have already adopted OS. It is those who see it and think ‘This is great! Sign me up!’
Diffusion of Innovation theory: reduce barriers
The rate of diffusion depends on how easy it is to overcome the barriers for adopting the idea.
Time
Barriers to think about while reaching the Early Adopters
“The Open Science/Open Source train has left the station.”
Next slides are what I didn’t get to during the ESIP talk
Need to change our messaging
Q2. How should we talk to leadership about the NMFS Openscapes mentor program?
Geoffrey Moore (1991) “Crossing the Chasm”
Time
Early Adopters
This is great! This is going to make our work so much more robust!!
Check out this cool new R package!
Wow, that is so cool. I want to try that!
I have a VISION of what our workflow would look like if everyone embraced this!
GitHub Actions! Wow! Let me try it too! 12 hours later…. That was so much fun! I learned so much!!
Early Adopters
Visionaries
Can see the future potential given the ideas and a test example
Looking for fundamental breakthroughs AND ARE EXCITED BY FUNDAMENTAL BREAKTHROUGHS
Willing (and excited) to put in substantial effort to get new technology working
Have the temperament to turn insight of potential into a working project
Risk Takers
There can be a disconnect between the vision in their heads and what they actually have in hand
Too willing to put in substantial effort to get new technology working?
Not all that excited by incremental change
Early Adopters need to convince ‘Early Majority’
The Chasm blocking organizational adoption of new ideas
Early Majority: Early Adopters need to target this group
Open to new technology but will need convincing
Risk averse; don’t want to head down a ‘rabbit hole’
Price sensitive
Note in many organizations the top leadership is not in the “Early Majority”. Yet the Early Majority are those open to the new ideas and are the ones the Late Majority will be influenced by.
Pragmatic
Want to see that the idea works and has value
Late Majority is next group after Early Majority and often includes organizational leadership
Generally resistant to change
Risk averse, price sensitive, really don’t want staff going down ‘rabbitholes’
Will need to see strong numbers to convince them: 1) Other organizations that have adopted this, 2) Fraction of people who have adopted, 3) “New” is a negative for this group.
Leadership of the organization is more likely to be in this group.
How to Cross the Divide
The Chasm blocking organizational adoption of new ideas
Key points