Considerations For Ethical And Responsible Product Experimentation
The mantra for product experimentation should always be “experiment with your customers, not on your customers”. Always treat your customers with dignity and respect. Your ethics in business is all you’ve got. Don’t be evil.
You may also enjoy these articles:
Introduction
I’m often asked in workshops and presentations “What’s your thoughts on the ethics of experimentation?” This is an interesting question, and one that can be quite polarising.
I remember a time when presenting at one of Australia’s leading universities and the question of experimentation ethics was raised in the audience. The discussion in the audience got heated quickly, being quite intense, producing a divide, and effectively splitting the room.
There were those in the room whose position was “how else are you meant to innovate if you can’t experiment?”. Let’s call them … The let’s just get on with things, make progress and learn as fast as we can crowd.
And there were those that were diametrically opposed.
Let’s call them … The “you shouldn’t be running any experiments without getting approval from an ethics board crowd”.
What split the crowd? The sectors they worked in.
Some people are passionate about the ethics of experimentation, particularly when they are new to experimentation and curious how it all works.
Ethics and morals are important in business. It’s all you’ve got to hang your hat on. If it doesn’t feel right, then you probably shouldn’t be doing it.
In this article we’ll discuss the following:
Ethical principles of social science research
What matters most is how you experiment
Why experimentation is necessary to learn?
Experimentation with vulnerable populations
Aren’t you just tricking customers?
Principles for ethical and responsible product experimentation
1. Ethical principles of social science research
One thing that product experimentation and clinical research has in common is the language that they speak.
They both speak the language of experimentation.
We’re both intent on understanding human behaviours better so that we can improve people’s lives.
“Like clinical scientists, product managers formulate hypotheses, design, and conduct experiments, gather data and information, interpret results, form conclusions, and communicate results”
However, there aren’t any research guidelines or ethical principles for product experimentation.
There’s no requirement for product managers to attend a review board or ethics committee prior to executing experiments.
Scientific study review boards became a necessity to prevent infamous medical research such as Phillip Zimbardo’s Stanford Prison Experiment or Josef Mengele’s human experimentation at Auschwitz prison camp.
For product experimentation, it more comes down to what a business is comfortable with, and the desired risk appetite.
Product experimentation is low risk on most fronts. The human cost, and human stakes, are significantly lower than clinical research.
I thought that it would be interesting to understand the guiding principles for social science experiments, seeing if there are any overlaps that could be applied to product.
Here’s what I discovered …
There are typically seven fundamental principles that should be considered before starting a social science research project:
Respect for persons
Beneficence and non-maleficence
Justice
Informed consent
Confidentiality and data protection
Integrity
Conflict of interest
Now, let’s look at each of these principles at a high-level.
1. Respect for persons
Summary:
All participants in research must take part voluntarily, free from any coercion or undue influence, and their rights, dignity and autonomy should be respected and appropriately protected.
Takeaways for product experimentation:
Applicable. Customers should always be treated respectfully, with dignity and empathy when conducting research. In product experiments, customers don’t opt in for obvious reasons. Typically, customers don’t even know they’re participating in an experiment.
2. Beneficence and non-maleficence
Summary:
Research should be worthwhile and provide value that outweighs any risk or harm. Researchers should aim to maximise the benefit of the research and minimise potential risk of harm to participants and researchers.
Takeaways for product experimentation:
Applicable. Executing experiments, for experiments sake, without clear learning objectives, is a waste of time. Weigh up the risk-benefit of any experiments.
3. Justice
Summary:
Researchers need to consider the overall societal impact of their research both in the selection of participants and the benefits and burdens arising from it. Research should be just as between different members or groups in society.
Takeaways for product experimentation:
Applicable. Customer sampling for product experiments should always be representative for the business problem being investigated.
4. Informed consent
Summary:
Informed consent requires that research participants should be given appropriate (a) information about the research (b) in a comprehensible manner (c) without duress or inappropriate inducement.
Takeaways for product experimentation:
Not applicable. We don’t want customers to be knowingly volunteering or self-selecting for product experiments. The experiment is immediately invalidated as customers could change their actions and behaviours if they know they are participating in an experiment.
5. Confidentiality and data protection
Summary:
Individual research participant and group preferences regarding anonymity should be respected and participant requirements concerning the confidential nature of information and personal data should be respected.
Takeaways for product experimentation:
Applicable. Say no more.
6. Integrity
Summary:
Research should be designed, reviewed, and undertaken to ensure recognised standards of integrity are met, and quality and transparency are assured.
Takeaways for product experimentation:
Applicable. Report and communicate the results of experiments objectively. Don’t cherry pick the data, put spin on the data or create a narrative that doesn’t exist to meet ulterior motives.
7. Conflict of interest
Summary:
Researchers must disclose anything that may be perceived by others as a potential conflict of interest. The independence of the research should be clear.
Takeaways for product experimentation:
Applicable. Report and communicate the results of experiments objectively. Don’t cherry pick the data, or spin the data, to create a narrative that doesn’t exist, or support ulterior motives.
While it’s possible to draw some parallels between the ethical research principles used for social science experiments and product experimentation, the context is completely different.
Product experimentation is not the same as social science experimentation.
They need to be approached differently.
Should product experimentation be treated the same as medical research?
No.
2. What matters most is how you experiment
Experiment with your customers
My mantra for experimentation is always “you’re experimenting with your customers, not on your customers”.
That’s an important thing to acknowledge.
“Product Managers have an obligation and responsibility to treat all customers with respect”
It ensures that you’re always partnering intimately with your customers to understand their pain, needs, problems and JTBD to solve their problems more effectively and make their world a better place.
You’re constantly learning in tandem with your customers so that you can deliver better solutions and experiences.
The more successful your customers are, the more successful your company becomes.
There’s mutual benefit in this situation.
It’s important to note that product experimentation is not social science experimentation. They’re not one and the same.
They shouldn’t be treated the same.
Social science experiments, when compromising on ethical grounds, can do harm to experiment participants.
“Bad social science ethics can be inhumane. Psychological, physical, and social harm can result for participants”
Run of the mill product experiments, at worst, can be a little bit annoying, providing a bad user experience.
There’s a big difference.
Minor inconvenience of a product experiment versus physical, psychological, or social harm of a social science experiment?
I argue that product experimentation shouldn’t be compartmentalised with social science experimentation.
They’re not comparable relatively.
Manipulating Facebook users for science
On the flip side, Facebook copped a lot of flak after running an experiment on its users to manipulate their emotional state.
The experiment aimed to investigate whether emotions are contagious via online social networks.
For one week, Facebook showed people fewer positive or negative posts to people in News Feed, and then measured how many positive or negative words they included in their own posts.
“Data scientists at Facebook manipulated the News Feed of 689,003 users, removing either all the positive posts or all the negative posts to see how it affected mood state”
When this experiment escaped into the public domain the optics were not good. Playing mind games with users in the “interests of science” is a bad look.
I’m sure the Facebook Emotionally Manipulates Users Study wouldn’t have passed the ethics board.
Interestingly, this study was designed by Cornell University but executed by Facebook, bypassing the Cornell University ethics board.
“This experiment certainly raises the issue of where to draw the line on A/B testing and product experimentation more broadly”
A case of experimenting “on” your users, not “with” your users.
While the results of the experiment were negligible, it’s the ethics of the experiment that were questionable.
Mean number of positive (Upper) and negative (Lower) emotion words (percent) generated people, by condition.
Where does this end?
In this situation it’s reasonable to suggest that Facebook acted unethically.
This wasn’t your typical, everyday product experiment. There was a deliberate attempt to manipulate user emotions and feelings.
But where does this end?
You can easily end up down a worm hole very quickly.
“Every brand on the planet, in some way or another, is trying to manipulate our emotions to purchase products”
It doesn’t matter if it’s a product, brand, service, charity, politician, or social cause, they’re all running A/B tests to discover what’s most effective at influencing our behaviour.
We don’t purchase products based on linear, rationale thinking. Much of our purchasing behaviour is driven my motivations.
Coca Cola’s tagline is “Open Happiness”. You’re meant to feel happier by opening a can of Coca Cola.
“Is Coca Cola any less unethical than Facebook, influencing people to guzzle a can of Coke - a drink which has zero nutritional benefit, potentially causing obesity and diabetes with long-term consumption”
The difference is that someone can always opt out of drinking a can of Coke, whereas Facebook users weren’t aware they were unknowingly participating in an experiment.
3. How else are Product Managers meant to learn?
Without experimentation, it’s very difficult to objectively quantify and assess the level of risk in launching your new product or feature.
But how else are product managers meant to learn you ask.
What we know with behavioural science and economic theory is that it’s like looking through a broken set of binoculars.
Neither market research into customer behaviours or economic theory provide a complete view of human behaviour and motivations.
In the words of advertising tycoon David Ogilvy …
“The trouble with market research is that people don’t think what they feel, they don’t say what they think, and they don’t do what they say”
And trying to use economic theory (E.g., supply and demand) to understand customer motivations and behaviours, it’s just too theoretical, rationale and linear.
Economic theory doesn’t help you to understand ‘what’ your customers do and ‘why’ they do it.
So, what on earth are Product Managers meant to do?
Experimentation is the only reliable way of testing your big idea.
“Experimentation is the only way that you can understand if what customers are telling you is true”
Qualitative research guides insight generation and idea generation. Quantitative research guides product decision-making.
It’s risky to launch your new product based on qualitative customer insight alone.
Experimentation is a necessity to validate the customer say do gap.
4. Experimentation with vulnerable populations
One of the scenarios where you need to tread a little bit more carefully is when experimenting with vulnerable populations.
There can be situations where sensitivities need to be applied – customers in distress or under conditions of hardship.
Sometimes, these situations can present an opportunity for businesses. For instance, to support their customers with new, novel solutions in time of a crisis (E.g., Natural disaster relief)
“Vulnerable segments can present an opportunity. They’re often an under-served population and ripe for innovation”
It can be satisfying to support vulnerable customers to save money and live a better life. Every little bit helps.
Depending on the type of experiments that you’re running, it’s often good to view experimentation as no different to running an outbound Marketing campaign.
The same data governance rules that apply to Marketing, also apply to experimentation - don’t contact customers on hardship programs, bad debt, overdue debt, do not market lists, do not contact lists etc.
Always respect your customers.
To paraphrase Google, don’t be evil.
5. Aren’t you just tricking customers?
I get asked this question a lot, particularly when referencing Fake Door experiments.
Aren’t you just tricking people? Well, just disappointing them (a tiny bit).
Without doing this, it’s difficult for Product Managers to learn about customer demand, interests, and preferences.
Otherwise, you’re just relying on subjective feedback and opinions.
The intent is not to be purposefully misleading or deceptive. That’s not good. It can also get you into trouble with the law.
I don’t advocate designing “Buy Now” experiments if the customer can’t purchase the product, and the service/product can not be fulfilled.
I feel that this oversteps the mark.
“When I run Fake Door experiments, I always run the experiment in a closed loop feedback system. The customer is never left wondering what’s next”
After the customer completes the desired action or behaviour that you’re measuring, don’t leave them hanging.
Use a pop up to advise the customer that the process is complete … “Oops we’re really sorry, we’re still in the process of developing this service. Stay tuned”.
End.
It’s important to manage the customer experience appropriately with these types of experiments.
A “button to nowhere” or a 404 error is not cool. It’s old school, resulting in unnecessary frustration and friction for the user.
It’s just a bad user experience.
6. Guiding principles for ethical and responsible product experimentation
How can you be a responsible product manager? Consider the following …
Ethics – with no ethics boards or committees to review experiments, it’s up to you as a Product Manager to design intentional and responsible experiments
Respect – always treat your customers with respect and dignity. Experiment with your customers, not on your customers
Conscience – if you’ve got a nagging feeling in your conscience that the experiment crosses the line, you’re probably right
Consult – get a second opinion from your product team, peers, or colleagues
Clarify – if you think that there may be legal implications and business risk, take the time to clarify your experiment and learning objectives with your Legal, Risk and Regulatory teams
Awareness – be aware of sensitivities that may exist when conducting experiments with vulnerable populations
Experience – don’t leave the customer wondering. Always close the loop on experiments to ensure a satisfactory customer experience
Implications – what are the implications of your experiment if things go wrong? Do a mini pre-mortem before you hit the launch button if you’re concerned
Conclusion
Product experimentation is not the same as social science research.
The ethics of product experimentation should not be treated the same as medical research.
The stakes are not the same.
Social science experiments can cause irreparable physical and psychological harm to participants if compromising on ethical principles.
Product experiments should always aim to treat customers with dignity, empathy, and respect.
The mantra for product experiments is simple, “experiment with your customers, not on your customers”.
If you’ve got reservations about your experiment, take note of your conscience. Listen to your moral compass - you probably shouldn’t be doing it.
Your ethics and morals are all you’ve got in business.
Don’t be evil.
Need help with your next experiment?
Whether you’ve never run an experiment before, or you’ve run hundreds, I’m passionate about coaching people to run more effective experiments.
Are you struggling with experimentation in any way?
Let’s talk, and I’ll help you.
References:
Before you finish...
Did you find this article helpful? Before you go, it would be great if you could help us by completing the actions below.
By joining the First Principles community, you’ll never miss an article.
Help make our longform articles stronger. It only takes a minute to complete the NPS survey for this article. Your feedback helps me make each article better.
Or share this article by simply copy and pasting the URL from your browser.