I'm currently doing research on optimal service pricing for a sales group using time series data.

They haven't done much formal data analysis before, but have ran some correlations on time series data that has not had its trend removed.

The result is that they are considering making management decision on this faulty data because the data has perfect correlation between the two variables when really there is no statistical relationship at all if you account for time and trend. It also turns out my boss did some of the analysis himself.

How do I explain my objections to this "data driven" analysis without completely undermining his work?

  • Use phrases like "Please correct me if I'm wrong" or "Can you spot any mistakes I've made in my findings?" to ease the undermining tension. If he's a good boss he won't take your findings personal and correct an error.
    – Mike
    Dec 7, 2017 at 22:27
  • Did your boss assign the research you're working on? If so, what was his motivation? If he is smart, he might ask you to verify that his conclusion is correct before making decisions based on it. Dec 8, 2017 at 0:51
  • Would someone in your position be expected to know more about this than your boss?
    – user8365
    Dec 8, 2017 at 16:02
  • @AffableAmbler he assigned me to work on the project. Till now they haven't had anyone official to work on the project so he was doing it since its his business.
    – EconJohn
    Dec 8, 2017 at 16:48
  • 1
    @EconJohn If he asked you to work on this, don't be afraid to share your findings, even if they go against what he thinks he knows. I know it can be frustrating to use stats to try and explain to non-statisticians why their findings are wrong. Just make the best case you can and avoid being condescending. Remember why you're both there--to help the company make money. Dec 9, 2017 at 6:09

3 Answers 3


How do I explain my objections to this "data driven" analysis without completely undermining his work?

Probably you are overthinking this a bit. You were hired for a purpose, and it is your job to do such analysis and report your finding and concerns about it.

I see no problem with you asking your manager for a one-on-one meeting, where you can explain to him your findings on this time series analysis you have made, and how they contrast with the results obtained with the previous method.

After that, you can then hint to him " I fear that if we are not careful this might end up bad for us. How do you consider we should proceed?"

This way you are respectfully exposing the facts that you see, and your worries, and giving him the chance to take managerial decisions based on that. If he decides to ignore this then that would be his call, and few you can do to make him take the choice you want. It is best if you stick to the facts and results from your analysis, and let your manager deal with the managerial decisions.

  • 1
    Your first paragraph after the quote is jumping to conclusions (we do not actually know if this was what the OP was hired for, and maybe he is overstepping his mandate with this kind of analysis), but +1 for pointing out that it's the managers prerogative to, well, manage. Dec 8, 2017 at 8:30
  • @EikePierstorff "I'm currently doing research on optimal service pricing for a sales group using time series data." - I think OP clearly stated here that it is his job to do that. Researches and finding should be reported to give his task meaning. OP is in position to obtain and report such results.
    – DarkCygnus
    Dec 9, 2017 at 18:44

They haven't done much formal data analysis before

It sounds like your job position is new. Is that a fair statement to make?

So they knew something worked before, and now they're tracking data and trying to make decisions based on that. It's fair to say that reality is they don't know exactly what to draw from this data. As such, if you want to make an impact, don't just say they made the wrong correlation because they believe they drawn their conclusion based on some trend they see. Ex "every summer, people start buying apples, so let's buy double apples!" Matter of fact, don't even bring it up. Instead, make a presentation on what you think is right, and explain it. Ex. "We should only increase our apple stocks by 10% because x,y, and z."

It could be a trial and error situation. It might be a good thing to see it fail then pick up where it fails driving home the point you originally made.


While being confident is imperative, don't convey each and every issue to your supervisor - be particular. "At the point when supervisors are barraged with relational issues between associates, they have an inclination that they are keeping an eye on than driving," Grenny says. "Continuously attempt to determine issues with your partners previously rushing to the manager." Bigger worries that effect your execution or the execution of the association are supervisor commendable, as are times when you have to claim up to a mix-u

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .