From 3-4AM always low paying clicks!

What would you do? Replace with YPN?

[dzcap]

I've been monitoring my hourly stats for an entire week so far. From what I see, it seems like I always get very low paying clicks between 3-4AM PST. The eCPM doesn't drop though because of a higher CTR during that time period.

What would you do to maximum profit during this hour? Replace with YPN? Optimize the ads more? What would you do?

[OptiRex]

I've been monitoring my hourly stats for an entire week so far.

Presumably you are referring to Adsense stats?

If so then I wouldn't take any notice of them, hourly stats are notoriously unreliable, about the only thing that is noticeable is when there have not been any stat updates.

To give you an example, if I were to go by Adsense hourly stats then the past hour has given me 40% of this morning's earnings after 6 hours...which I know is not correct, it's simply Adsense catching-up.

Some people swear there are accurate trackers available however I have never tried them since the possibility of gaming Adsense to increase one hour's earnings just isn't worth the effort.

IMHO your time would be better spent in adding more valuable and relevant content.

[dzcap]

The hourly earnings I track are fairly consistent, each hour is like exactly about 1/24 of earnings. Except the couple hours in between, such as 3-4AM where I noticed the low paying clicks.

I am really not that bothered by it, but I keep thinking of the possibilities if they were normal paying clicks instead. So I don't know, it would be a pain to replace that block during low paying hours only. But if I can gain an extra 10-20% just by doing that, I think it'd be a wise decision.

Or perhaps there are some tricks to get them to normal paying values instead?

[OptiRex]

each hour is like exactly about 1/24 of earnings.

Then, IMHO, you have THE most unusual site on the Net!

I'm sure someone will tell me I am writing bunkum however I know of no web site which has consistently equal 24 hr traffic periods.

You have the niche all of us are looking for...what is it? :-)

[dzcap]

Well I wouldn't really call it exactly 1/24 but very close. It is like the first 12 hours gets 60% of the earnings, and the last 12 hours gets 40%. Most of the time, the earnings each hour are consistent. So that was why I thought improving the click values during specific hours (3-4AM) would be very profitable.

[OptiRex]

So that was why I thought improving the click values during specific hours (3-4AM) would be very profitable.

I would have thought that it would have to be a potentially substantial sum to warrant this effort.

For instance, what do you estimate you earn at the moment for that specific period? USD10/50/100?

Could you increase it by 10/25/50%?

Let's say you're earning USD10 per hour, which is USD240 per day, a 10% increase for that hour would only yield USD1.00 and even a 50% increase USD5.00.

I would've thought a helluva lot of effort to gain possibly USD150 per month for a site which is earning USD7,000+.

IMHO anything less than these figures and you're wasting your time.

Even USD50 per hour with just a 10% increase would not be worth it however with a 25% increase then would the USD375 extra be worth it when you're already earning USD36,000 per month?

[dzcap]

The benefit is minimal at the moment, but I plan to use this strategy across all my sites, the benefits could be enormous. I thought this might be another good direction towards increasing revenue.

However, it definitely won't be a 10% or 25% increase, most likely single digits. After all, we're talking about just 1 hour here.

[hunderdown]

Another thing to consider--why assume that YPN would perform better than AdSense at 3-4 AM?

[elsewhen]

i agree with hunderdown... i think the click payouts are the lowest during that time because the big u.s. market is asleep. the people who are clicking are from other countries, and those that are advertising to those countries are probably bidding less per click.

i bet ypn would be the same, but that doesnt mean you shouldn't give it a try.

[Bluepixel]

Your EPCM difference probably correlates with different region visitors. I have the same thing, in the morning, EPCM is generally higher, but it slightly decreases as the day finishes.

[elfred]

I expressed a similar thought here: [webmasterworld.com ]. I have a lot of stats and I can predict what will happen with a high degree of reliability. When, on October, 17th I started getting bad figures, I predicted a 17% drop in earnings at the end of the month and now I can confirm that drop: 17%. I hope today (November, 1st) isn't representative of the whole month, otherwise I'm going to face a 66% drop this month. Something similar happened one year ago, but that was clearly due to AdSense changing niche. Now I assume something similar took place, but the difference is much more subtle and I can't figure out how to bring the niche exactly back where it was.

[david_uk]

I'm awake when the US is still asleep, and early in the morning here I tend to see the ads that are less well targetted, and income tends to pick up when the US is at it's corporate desk.

I personally doubt that changing to YPN for those hours will make a lot of difference, but anything is worth a try! Just monitor any expermient closely so that you know of it's worked or not.

[jomaxx]

I seriously doubt that it would be worthwhile to turn your ad serving inside out simply to take advantage of a 1-hour glitch.

Anyway, I also see a regular and rather pronounced variation in earnings each day. But in my opinion it's probably not "real". It could be an artifact of the algorithms Google uses to determine how much advertisers pay and how much you receive, or it could simply be that a percentage of clicks throughout the day are set aside for more detailed filtering and analysis that isn't done in real time.

[AlexMiles]

I'd do it if it could be done in an automated fashion and wouldn't take more than a couple of hours to arrange. You could have a script that looked at the time on the server and did mysterious stuff.

What exactly, I don't know. Swapped out Adsense for YPN or Chitika?

[sailorjwd]

I find that I get oddly high epc clicks during the first 4 hours of the day... may not get another click on some channels all day long. Or, if I do they are often worth 10-20% of those first clicks.

Probably a stat issue.

[21_blue]

dzcap wrote:
>I've been monitoring my hourly stats for an entire week so far. From
>what I see, it seems like I always get very low paying clicks between 3-4AM PST.

I don't think that a week is sufficient to draw conclusions from. However, we've been monitoring time-of-day stats on an occasional basis for about 9 months, and our data supports what you say to some extent, but in a slightly different time slot - income tends to slow down between 4:30 and 6:30am PST.

We put this down to the lunch period in the UK (plus a bit of stats reporting delay). The reduced CPC may be due to the shift in demographics of surfers away from UK business, which perhaps attracts a higher rate than topics associated with 'personal' surfers or those in other countries. Replacing ads with another scheme at that time may not make much of a difference, because the change is associated with market movements.

However, you'll never know for sure until you try it, but I recommend you collect much more data than a week, so you don't compare on statistically unreliable samples.

[21_blue]

sailorjwd wrote:
>I find that I get oddly high epc clicks during the first 4 hours of the day

Another possible explanation is that you have:
- one advertiser who has a relatively high bid and low daily budget
- another advertiser with a high bid and high budget
- the rest of the advertisers have lower bids and budgets

At the start of the day, the prices are driven up by competition between the two high bidders. But once the first advertiser exhausts the budget and drops out of the auction, this enables the second advertiser to win the auction with a much lower bid for the rest of the day.

[dzcap]

I am thinking of scheduling wget or some other automation to replace files at 3AM and back to the original at 4AM. I plan to do this for the 5 hours of poor eCPM. I agree a week's monitoring is probably not enough, but if I wait for a large enough sample, just think how much money would be lost during that time.

I find the first 2 hours of the day always pay the most for me, but after the 2 hours, eCPM drops over 70%. This has been consistent for over a month already..

[21_blue]

dzcap wrote:
>I am thinking of scheduling wget or some other automation to replace files at 3AM
>and back to the original at 4AM. I plan to do this for the 5 hours of poor eCPM.

The only potential spanner in the works for this could be stats delay - 3am to 4am may not be the lowest period for clicks, but simply the lowest period for reporting which might be influenced by things as innocuous as Adsense overnight runs (payment calculations, EFT, maintenance, etc.). You want to make sure you swap files when the income drops, not when the reporting might drop for other reasons.

If you have the ability to replace files so easily, can I suggest the following experiment:

• set up 24 custom channels to cover each hour of the day
  
  
• schedule wget to swap the files in automatically on each hour
  
  
• collect the data for a few days, and compare the performance of the channels to assess which hours perform best/worst
  

That way, when you choose to swap ads you can do it on the basis of the actual click-time performance rather than the (less reliable, potentially delayed) report time.

[sailorjwd]

Added info on my 0-4am high epc phenomenon:

Looking at this for a year.

Happens across many diverse channels

Not a delayed stats issue.

I thought adwords evens out the distribution of ads through the day. I believe the only way to thwart that system is to lower you budget in the middle of the day routinely (I think someone got kicked out for doing that).

Other issue: for me CPM declines throughout the day, every day (I believe this is tied to delayed impression counts and to the high EPC clicks in the morning).

[21_blue]

sailorjwd wrote:
>I thought adwords evens out the distribution of ads through the day

That's the theory, and perhaps it happens for some. But we use Adwords and our experience is different - we have been running out of budget part way through the day. Not every day (weekends are slow) but it happens most weekdays.

[OptiRex]

An old business friend of mine (nearly 80 yrs) once said to me: "Too many people waste too much time trying to save a little when they should be spending that same time more productively and actually making more. The increase in profits will more than offset the small savings made."

You are in danger here of focusing on something which is outside of your control since you are not privileged to the information on which to base your assumptions.

You feel that your stats are showing you the reality, but are they?

Will Google tell you? Nah...

If you read lammert and 21_blue's posts about standard deviation you would then understand why you're banging your head against a brick wall.

[webmasterworld.com...]

[webmasterworld.com...]

[dzcap]

I agree with you, very good advice.

I didn't idle at this problem. I've assigned somebody to analyze this already so I have the time to work on other projects.

[aeiouy]

You have to also consider that some advertisers can and do limit their inventory to specific times of the day. If you have a key advertiser or two who are prevelent most of the time not advertise in the middle of the night, then it could impact this as well.

[21_blue]

aeiouy wrote:
>You have to also consider that some advertisers can and do limit their
>inventory to specific times of the day.

Is that a feature reserved for premium advertisers? I don't think I have an Adwords option to select times of the day.