When communication needs more than just the facts

Data doesn’t always win

Changing people’s minds is hard. But we often feel like it should be simple. Gather the facts, share them in a civil way, and you’re done. People’s emotional responses may create obstacles, but that only means our data isn’t compelling or that they’re being illogical. We should continue gathering and presenting our facts until people see reason.

We want this assumption to be true, and sometimes it is. We respond well to logic in areas where we don’t have entrenched beliefs or where conflicting opinions don’t question fundamental assumptions of our self worth. These discussions go smoothly because they don’t threaten our identity — that fuzzy, amorphous ball of Beliefs, Values, and Important Stuff™ that tells us who we are at our core.

In issues where our identity is involved, there’s a lot of evidence to suggest that logic cannot change someone’s mind, no matter how clearly stated. In these discussions, people’s decision making processes are structured more like a CS101 midterm. Nothing is well named, interfaces are murky, encapsulation is an accident (if it exists at all), and most things work by side effect. We did start building our mental code base when we were toddlers, after all.

Working on the web platform team at Chrome, we often find ourselves in emotionally charged discussions. It’s important that we are able to detect when conversations involve identity and understand the mindset people have when their identity feels threatened. Only then can we effectively communicate.

Detecting identity

Recognizing a conflict that’s touching on identity requires asking a few questions:

  • Is there disagreement despite the presence of seemingly unbiased, hard evidence supporting one side?
  • Is someone engaging in meta arguments, placing inordinate weight on technicalities or how things are said?
  • Is someone doing something that seems to be directly against their best interests?

Once you recognize that someone’s identity is involved, you need to understand what part of that identity feels attacked and what part of your argument is attacking it. There is no hard rule on what is part of someone’s identity. Ultimately, you need to exercise judgment and ask yourself what is important to the other person. However, there are a few common themes that you can consider.

Self worth

We all want the things we do to be valuable and will seek to disprove statements that claim otherwise. Someone’s self worth is likely to feel attacked when the topic of discussion is something that they have created. If you’re critiquing someone’s product, performing a code review, or commenting on a proposal and you detect identity, self worth is likely involved.


The easiest way to detect when someone’s values are attacked is to look for generalized or abstract responses to specific points that you make. For example, if you say “Chrome should build special affordances for Website X”, there are two kinds of negative responses:

  • “We’ve found that websites come and go often enough that targeting one specifically isn’t effective in the long-run.”
  • “That wouldn’t be an open ecosystem.”

The first response is open to a data driven discussion while the latter is someone’s value. You’re also probably touching on values if the individual is responding with moral arguments.

In group, out group

Humans are hardwired to categorize themselves and others into groups. If you detect that someone’s identity is being attacked in a discussion that involves groupings (users vs. developers, eng vs. design vs. pm, team A vs. team B etc.) it is very likely that someone identifies with one of the groups being discussed and feels the need to defend it.

Addressing identity

In identity-driven discussions, one or more parties have shifted their goal from “finding the right answer” to “proving you wrong”. Your argument is seen as an existential threat to their sense of self. Your data is just a weapon that needs to be invalidated. It is no longer input to be discussed.

At this point, to move forward, you must accept that logic will not play the only part in this discussion. Until you find a way to frame the problem that agrees with or empowers the other side’s identity, no forward progress will be made. This is true even if you are right and have concrete evidence to that effect. How you frame your comments will matter tremendously.

Once you know that identity is involved, you can construct a plan forward. Again, the final structure of your plan will require judgment, but there are a few general strategies you can use depending on how identity is involved.

Self worth

The key here is to focus criticism on the outcome rather than on the person. In some ways, this is obvious. You should never attack someone based on an idea they’ve proposed (e.g. “Only an idiot could have proposed this.”). But you must also keep in mind that people’s ideas are frequently treated as proxies for themselves. Saying “this idea is bad” is often equivalent to “this person is bad”. Instead, frame critiques in terms of the outcomes:

  • Will this achieve the growth we’re looking for?
  • This might lead to consequence Y. Are we comfortable with that?

Also notice that the critiques above were stated as questions. Unless someone has proven themselves to be combative, it is good to give them a chance to correct an error on their own by answering your question and updating the proposal.

If you must provide a direct critique of an idea, preface it with positive feedback first. Make sure to include how someone who was intelligent, competent, or strong could have reached an similarly incorrect answer: “This area can be really counterintuitive. I messed it up a bunch until I had been on the team for a while, but I’ve found X is generally better than Y.”

Keep in mind, criticism goes a long way — you don’t have to point out every flaw. Harvard Business Review found that a ratio of 6:1 for positive:negative feedback produced the most effective teams.


Working around values can be tricky. It may be worth considering if there is a compromise you can make up front to show a person that you’re considerate of their values. Alternatively, try to reframe the problem in a way that puts it in the context of an alternative value that is more important. To continue our example from above: “I know that providing preferential treatment to Website X hurts the open web, but if we do it we could reduce data consumption in emerging markets and save people money they need to support themselves.”

In group, out group

The easiest way to avoid group identities is to speak in terms of things, not groups. Instead of saying “this is good for developers and bad for users” say “this will lower development time but will increase input latency by X%”.

If you do need to refer to groups, one Simple Trick™ is to replace any instance of “I” with “we”. This can override any passive group identifications and enforce that the primary group is you and the other party working together to solve the problem.

When all else fails

Sometimes, it’s possible that identity is inherently entangled with the problem and no amount of rephrasing will help. In these cases, you need to find a way to reframe the other side’s identity to make it amenable to collaboration. Give the other side a way to accept your argument without feeling like they’ve compromised their values. Redefining someone as “always quick to adapt and find new solutions” rather than “always right” may make them more amenable to discussing a current flaw.


Even with a carefully crafted message, resolving identity-driven conflicts will always be hard. Emotions will be involved. But if we accept that the discussion will require more than just data and if we can frame the discussion in a way that empowers everyone’s identity, we can find a path forward that leverages the strengths of everyone.