HN Explorer

Why you should ALWAYS CC someone

98 points by dsirijus ↗ HN
I ran a little statistics experiment over my Gmail inbox and here are the results:

Mails that I haven't gotten replies to and mails I did have, are in ratio ~1/13. In those mails not replied to, mails WITH someone in Cc and mails WITHOUT anyone in Cc are in ratio of ~2/11.

Now, this sounded all obvious - the more recipients you address your e-mail, the higher are the chances you will get a reply, right? Right, but not quite. :)

For the last 2 months, I've created a imaginary CEO assistant e-mail account named Alex. I've Cc-ed Alex to all my mails and Alex has not directed any mails to anyone, nor received any as To recepient. Select few contacts were exempt from the experiment, but that's taken into statistical account. So, Alex has basically been completely non-interactive in all this, besides being in Cc.

After two months, mails that I haven't gotten replies to and mails I did have, are now in ratio ~1/22.

I have my own explanation of the phenomena, but I'd like to hear yours.

:{-

69 comments

[ 5.8 ms ] story [ 134 ms ] thread
Sword_Monkey
If someone else is cc'ed in then people feel pressure to reply as they aren't just ignoring one person, there's a whole other person expecting a reply. Even if that person is someone they've never met, or know of.
dsirijus
I don't think that alone can account for doubling the response rate.
Michiel
Why not?
dsirijus
Because in most of those mails it was mission critical that I receive the response. The process grinds to a halt if I don't receive response, irregardless of anyone and anything else. Both Alex and I were tied to the same process and action. The priority of it is unchanged.
csallen
This doesn't mean that recipients can't feel additional pressure to respond due to the presence of another party.
krmmalik
I haven't fully understood what you're trying to say, although i am convinced you might be onto something.

Are you saying that emails that you send to other recipients have a higher chance they'll be replied to if you CC Alex on the outgoing email?

can you tell me what the increase in replies is as a percentage?

dsirijus
x2. I couldn't believe it.

I ran the statistics just by plain labeling things consistently, but they're pretty solid.

EDIT: almost x2

philbarr
Perhaps, over the last two months you've been running the experiment, you've been thinking about actually wanting a response and have automatically been constructing better emails that people feel they can reply to. Unlike the "brain-dump" ones I sometimes get that just don't deserve a response.

To be perfectly honest, if there's more than one person on the email, I'm more likely to hope it's not my problem...

dsirijus
Well, I've always tried to construct my e-mails to the best of my ability. I didn't noticed any change in my content - it has all been pretty standard.
tg3
Unfortunately you not noticing any change in the content doesn't make for very reliable or convincing statistical results. Thus the suggestions for making the experiment blind to you by choosing whether or not to CC after you compose the email.
dcminter
Ack. Always address one of the recipients by name in the body of the email to ensure that it's not SEP'ed.
adityar
isn't there a name for that? something like diffusion of responsibility. the same reason we don't help one person in trouble when there's more than one potential helper...
dsirijus
But this technically isn't diffusion. It's concentration of responsibility.
ekanes
Yes, but there's a critical difference. In this case there's a clear recipient - someone asked to do something/reply. In the case of the Genovese effect, no-one is specifically asked to do anything, and no-one does.
eric-hu
So to design a better controlled experiment, the OP should write a statistically significant number of emails (>100...maybe 1000+) without sending them. An app should perform a double-blind study by randomly choosing half those emails to go to multiple recipients (all but one being fake) and choosing the other half to go to one recipient only.

Anyone see any holes in this approach?

Tyrannosaurs
How about the idea of writing 100+ e-mails without sending them?

You'd be better off having a mail plug-in randomly selected whether to add a cc at the time of sending rather than hoarding them.

Your mechanism is valid as a trial but not particularly workable in the real world.

You might want to also control for factors such whether the mail is internal or external (assuming it's a company), message length (longer messages might be more or less likely to be properly read and responded to), message importance (you'd expect more replies to higher importance messages), attachments and so on.

mcherm
"without sending them" is unnecessary. Modify your email client to randomly choose (AFTER you write your email) whether or not to CC Alex. Or just flip a coin after writing it and before hitting send.
TobbenTM
I would think it is because the email becomes more personal and social. More people 'collaborate'.
jtheory
In particular, more social.

My suspicion is that when recipients see the cc with a name they don't know, they're suddenly in the position of making a first impression on someone that don't know yet, but will probably be working with in the future.

Another possible (related) thought -- here's an assistant who will be taking over some of these conversations, and thus they're probably talking about me (...so if I don't reply now, that's what they'll be discussing).

People care what others think of them; these people already know the original sender and have a stable relationship they don't feel will be threatened by a slow response (which easily slips into no response); but this adds a new person into the mix.

3xBcF32xNi3kh7i
Would you mind sharing the statistical sample size? It's hard to talk about statistics without it.
dsirijus
It was not rigorous, but here they are...

Pre-Alex I've ran through 5 months of mails, which amounted to ~5k mails, and sorted them out (there was already label Waiting done for me). Post-Alex, in 2 months of mails the sample size was ~2k mails.

Number of e-mail addresses contacted within those 7 month period is ~70.

TeeOff
So you send ~50 mail every (working) day? Could your sheer volume skew the result? (as in: spam people and they start to ignore you, put someone on CC and they have reason to believe it's worth reading?)
dsirijus
I send mails every (week) day, and favor it in front of any other form of business communication.

I've sent 17 since I've started this thread.

citricsquid
I'm curious, what's the average email of those 17? Like, what's the subject and/or value to the recipient and you?
dsirijus
Sending stuff for approval and approving stuff done. This is a busy day, though.
adityar
Is the content being mailed the same? are you mailing the same people again before and after alex was created? are you split-testing -- 1/2 recipients get the email without alex in CC? You might need some more experiments to attribute it to pure CC.
dsirijus
Content is the same, but that's obviously subjective. The conversations were either all with alex or without, meaning, no reply requests were sent to the pre-alex mail with alex in cc and I haven't split tested.
adityar
how about the recipients? If they had an email from you before (part of the 5K set), then might be more open to replying to the second one (in the 2k set) even if Alex was not on cc out of guilt?
dsirijus
Most of the recipients are contacted on relatively regular basis (some daily, some weekly, some monthly).
israelyc
Really interesting, have you tested the effect on gmail priority inbox? it could explain a higher open rate.. and it would awesome to know that cc is part of their rankings..

[EDIT]- Also, if you all set with labels already, would you mind continuing the test with something like SigBuzz (g apps) to track open rates and keep us posted?

dsirijus
No, I don't use priority inbox.
mobweb
I think he means that it might have an effect on other people's priority inbox. It could be that if an incoming E-Mail has another recipient, GMail thinks this E-Mail might be more important and place it in the priority inbox.
dsirijus
Yes. I misunderstood.

I know for a fact that most of those contacted do not use Gmail. Only one confirmedly does.

There might be other corporate non-public solutions that do that, on the other hand.

dsirijus
Well, it's gone public now, it's tainted. I'm already receiving "You sleazy bastard. :)" mails.
jmmcd
> I've created a imaginary CEO assistant e-mail account named Alex

What's Alex's email address? If it's "Alex Q Alex, CEO <alex@your_company.com>" people will think the email is more important than if it's "alex2009@yahoo.com". Unless of course they work for or know your company, in which case they'll know the CEO isn't him. Actually I'm trying hard to think of a scenario in which the recipients don't just think you mistyped the CC, in that case.

EDIT Urgh, you said CEO assistant. I can't read. Disregard this.

dsirijus
Alex was not CEO, he was CEO assistant, but that doesn't matter at all since no one actually sent mails from that address, and subsequently have not read the his mail sig.

The adress was alex@your_company.com

Praveens
If we are to neglect the importance of the email and solely judge on one's replying instinct when there is a email id in the cc field, then i feel i am obligated to reply as there is a certain amount of importance attached to the content of the email and its reply would be monitored by not just the sender but also the person in cc.
n3x
Very cool experiment, kudos, and i can see why it works. But did you actually say in the emails "ccing my assistant?"

Even if you didn't i am not sure how legal it would be. Just by ccing someone, you mislead the other party, eg that you have someone else in your company. Of course the argument for "don't use for important emails" is invalid, why you can mislead some and not others?

And when in doubt, i say no.

dsirijus
No, no e-mail sent nor received had any mention of Alex in any way outside him being in Cc.

I hardly believe it's at all illegal to have an e-mail address in Cc field, even if it's imaginary. E-mail isn't Facebook. :)

udp
> But did you actually say in the emails "ccing my assistant?"

They'd be able to see that anyway. It's not the same as BCC.

oskarth
You should try different names and titles for the assistant as well:

- Instead of Alex, what about Sarah (?)?

- Instead of a standard name, what about a noble rank / recognizable surname (+?)?

- Instead of CEO assistant, what about personal assistant (-?), accountant (+?), future candidate of CEO (+?)

What about varying the person(s) involved depending on the type of project? If it's a issue of numbers, involve some accountant. If it's about coding, involve a senior coder, if it's about direction of the company, involve a imaginary future board member. The possibilites are endless - just make sure you don't get caught lest you want people to think you have some kind of disorder.

Just the idea of having someone else looking over your shoulder could increase the conscientiousness a lot and keep people on their toes. If this got to become the mainstream way of doing things in some subcommunity it would be horrible though.

gaius
Alex is a good name for this, as it's gender-nonspecific. I tend to assume Alex's are female unless I know otherwise (like Lesley's or Hillary's).
caw
I wonder what the M:F ratio of the name "Alex" is. I agree that it's not really gender specific, but I only know 1 "Alexandria", and she doesn't go by "Alex". This could be a function of my university (70:30 M:F), or it could be generational (I'm in my 20's). As a result, I tend to default to male.

Perhaps different demographics respond better to different names, based on their preconceived notions of class and status. Does a really exotic name change anything?

oskarth
Right, didn't think of that. Here in Sweden Alex is just shorthand for Alexander (which means something like "defender of the men"). The female version here is either Alexandra or Alexis.
chris123
Chris if another option in the gender-nonspecific department.
helen842000
This is interesting because you're saying recipients are opening the mail and then choosing whether or not to reply after they read it. This shows e-mails are not just sitting unread.

I always figured that it was the subject line that mattered most and that once a mail was opened it was more likely they would reply.

I've experimented with improving e-mail structure. I always get a more detailed reply if I group any questions I want answering together as the final part of the mail.

[dead] sfgfdhgfdshdhhd
it all makes sense now. sincerely, alex@gmail.com
bad_user
For your statistic to be relevant, the experiment has to be blind: You must not know to which recipients you're sending the CC-enhanced version.

Then, the name of the CC-ed recipient may be important, so you have to do this experiment with multiple "imaginary assistant emails", trying out different variations of names (male/female, first-name/full-name).

liljimmytables
So, assuming one was a lazy slacker and wanted all their email to fall into a black hole for the purpose of work avoidance, have you got any advice other than "always mail people without CC?"

It's not for me you understand, it's for a friend.

dsirijus
Yes.

Never reply-to-all.

No subject.

Tyrannosaurs
Has anything about your role, the work you're doing or the success or otherwise of your company changed in the past two months?
dsirijus
Of course it has. Not significantly.
rubyrescue
i have a client who was a small 2-person startup before they got acquired. They were both doing client facing calls and emails with big multinationals trying to sell a large product inside a typical 100-200 person division. So they created a few fake personas in order to make it look like they could afford a support staff. It's fun to reply all and know that one of the cc'ed people doesn't exist.
oskarth
That reminds me of Julian Assange in the beginning of Wikileaks, making up a lot of fancy titles for imaginary personas in order to make WL seem a lot bigger than it really was.
quadhome
[citation needed]
oskarth
I don't remember where I read it, and I didn't manage to find it after a couple of minutes of googling. It could be a fabrication but I have two memories about it that speaks against it:

1) I think it was in a interview with JA himself (could have been the DDB book though which does indeed make it less credible)

2) It was said in a jokingly and positive way, which suggests it's not disinformation spread by someone trying to get them.

drx
You should re-run the experiment and make it double-blind.
dsirijus
My sample is tainted now. Someone else should do it.
scoot
In those mails not replied to, mails WITH someone in Cc and mails WITHOUT anyone in Cc are in ratio of ~2/11.

Am I missing something, or does this have no statistical relationship to the number of replies received?

Mails replied to could have all had a CC, or none had a CC and the CC ration for unreplied mails still be the same, no?

dsirijus
Yes, quite right. It's superfluous.
franze
the raw data would be nice - the full emails would be best,but just the sent w. cc,sent w.o. cc figures, time of message, time of response, number of reponse asking who is alex, .... anonymified recipents,... such data would be cool,
ArekDymalski
Pretty interesting. Before jumping to conclusions let me ask how many variables (time of day, wording of subject, frequency of sent emails, history of relationship with the recipient etc.) can you claim to "have under control"? Did you start CCing Alex for all 70 recpients? I'm asking coz there might be several explanations for the effect (assuming it's statistically significant and methodologically acceptable)
JoeAltmaier
Are you cold-emailing folks? In a close work group, I'd have guess replies would be near 100%.