1
$\begingroup$

Let's say I have a (ads live days, revenue) data set.
The data set shows how much revenues each ads generates during the days it is live.

  • ads1 generates 100 dollars during the 5 days when it is live.
  • ads2 generates 200 dollars during the 10 days when it is live.
  • ads3 generates 300 dollars during the 15 days when it is live.

...

so I have (100,5) , (200,10) , (300,15) .....
Imagine I have 100 ads in the data set and I can only launch 80 of them next month.
I have to decide which ads perform better so I can continue launching next month. What is the statistic model?

I was thinking using linear regression, to keep the nodes that have less variance .... but I am not sure if that is correct.

Improve this question
$\endgroup$
1
  • $\begingroup$ I would use the ratio of dollars / days for each ad and then rank the ads by descending order of dollars per day. Ads that make the most money will be in the top of the sort. Then pick the top 80. You are not trying to generate an average revenue per unit day (regression slope) for all the ads combined, but rather need to compare the ads with one another. $\endgroup$
    – user32398
    Dec 12, 2015 at 2:31

3 Answers 3

Reset to default
1
$\begingroup$

I am not sure to understand what you want to do because it does not seem to be a statistical problem to me.

With the only two variables you have, I guess the performance can be calculated as revenue/nb_of_days. And you'll choose the 80 more performing adds.

Why would you do a regression model ? What would be the predicted variable ? What would be the other variables ?

I guess your problem is more difficult than that but you need to provide more information.

Improve this answer
$\endgroup$
3
  • $\begingroup$ Thanks @Scratch ! Your solution definitely works. What if I want to predict if it is true the more days I am running a ads the more money I am making? So I will do for each ads. Ads A ( live for 1 day, revenue ) , ( live for 2 day , revenue ) , ( live for 3 day , revenue ) ..... and then I will get a slope A. For Ads B, I will get a slope B. If slope A > slope B , I pick ads A. Sorry, I don't have much knowledge about this, just trying to make a script to do auto pausing! $\endgroup$
    – peipei
    Dec 30, 2013 at 19:26
  • $\begingroup$ Sory, but I can't condone this. Revenue/# days just gives you the average earnings over the period the ad is up. It does not tell you which ad to choose and certainly does nothing to tell you about the future earning power of the ad. $\endgroup$
    – rocinante
    Dec 30, 2013 at 19:32
  • $\begingroup$ @rocinante I agree. I would not recommend doing that for a real world case. $\endgroup$
    – Scratch
    Dec 31, 2013 at 9:44
2
$\begingroup$

Why do you need any statistics? Statistical inference lets you generalize, but no generalization is needed here (if I understand you correctly).

What I'd do if I were you: Compute the dollars/day for each ad and pick the ones with the highest values.

Improve this answer
$\endgroup$
4
  • $\begingroup$ Thax! dollars/day does give me some meaningful information, but what if I want to predict if an ad will be continue performing good? I am just very curious is that a problem some statistical model can solve? $\endgroup$
    – peipei
    Dec 30, 2013 at 19:41
  • $\begingroup$ How is the average dollars earned per day a meaningful figure? The average earnings per day for ad A are not necessarily the same if ad A runs by itself versus the scenario that it runs along ad B and C (or any combination of the maximum 80 ads he/she can afford to buy). That's why I'm saying that this is a dynamic optimization problem. $\endgroup$
    – rocinante
    Dec 30, 2013 at 20:07
  • $\begingroup$ @rocinante. Given the information stated in the original question, I don't see what else you can do but pick the largest dollars per day. Given more information (sales on each day; reason for why some ads ran longer than others....), you can do more. $\endgroup$
    – Harvey Motulsky
    Dec 30, 2013 at 20:59
  • $\begingroup$ The information in the question was not really clear, so picking the maximum (or rather the top 80) is still wrong. How are those revenue figures calculated? Each ad running individually or some combination together? Peipei presumably knows the answer to this, as well as the costs involved. So he has the information necessary to perform a dynamic optimization model. There are many sources that estimate how much less effective ads become over time as well. This does not strike me as a lack of information problem. $\endgroup$
    – rocinante
    Dec 31, 2013 at 2:39
0
$\begingroup$

This is not a statistics problem. It is a dynamic optimization problem. You want to maximize your ad revenue, subject to the constraint that you only have X to spend and the placement of each ad costs Y.

Moreover, I think you need to revisit the assumption that the ads that are profitable in one month will continue to be profitable subsequent months. Aside from a few outliers, ads lose their effectiveness over time - which means that this is another constraint in your optimization problem.

If this is a real-life problem and not just a mental exercise, your best bet is to hire someone to do it. Kaggle, Innocentive, Elance are just some companies that have a database of reputable data scientist freelancers who can do this for you. (I don't represent any of the companies listed, and with some searching I think you can find others as well.)

Improve this answer
$\endgroup$
1
  • $\begingroup$ Thanks for the suggestion! I will check out those services. $\endgroup$
    – peipei
    Dec 30, 2013 at 19:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.