homepage Welcome to WebmasterWorld Guest from 54.237.38.30
register, free tools, login, search, pro membership, help, library, announcements, recent posts, open posts,
Become a Pro Member
Home / Forums Index / Google / Google AdSense
Forum Library, Charter, Moderators: incrediBILL & jatar k & martinibuster

Google AdSense Forum

This 70 message thread spans 3 pages: < < 70 ( 1 [2] 3 > >     
Repeatable, undeniable AdSense trend
The greater your number of clicks, the lower your EPC goes
NoLimits

5+ Year Member



 
Msg#: 11464 posted 10:12 pm on Jan 10, 2006 (gmt 0)

For the first year or so that I was with AdSense, I wasn't a dedicated webmaster. My site had a fair amount of traffic (1,yyy unique daily) but very little SERP activity.

My EPC for Adsense was the HIGHEST it's ever been during that time period.

I noticed a great opportunity to make money, and began to research SEO and continue the developement of my site. I managed to attain fairly good rankings in Google right off the bat, assumingly due to having a 4+ year old domain that has an active forum with lots of IBL's.

Suddenly, I had reached what I thought was eCommerce nirvana - with thousands of Google referrals coming in daily. It was a steady rise from 1,yyy to 1y,yyy uniques daily. My EPC fell, as my traffic rose - eventually hitting rock bottom at about 1/5th of what it was with less traffic.

I lived with this until YPN came out, and I quickly applied and was accepted early on. I gave Y! a good deal of my ad space because it was doing a little better for me on the bottom line. I cut adsense impressions down to the amount of views it was at prior to my search engine success - and magically over the span of a week or so EPC climbed back to where it was the last time my traffic (as far as AdSense can tell) was that low.

Then came Jagger -
NO search activity - none, virtually zero (from google)
I switched all my inventory space back to Google, because with the LOW CTR of YPN for me, it wasn't worth waiting for 100+ pages for that magical big dollar click - eeeking out 2,yyy uniques a day on good days, EPC remained fantastic with AdSense.

Then I met Google Images - traffic started to trickle in again, this time from images.google.com - I left AdSense with 95% of my ad inventory and watched the EPC fall day by day as my traffic increased.

This brings me to this week. Google images has been sending me tens of thousands of referrals - destroying any prior records by mid-morning. My EPC continues to fall...

The last 2 times I saw this happen - I dismissed it, after some complaining, as the result of depleted inventory due to my extra clicks - or smart pricing.

I will not dismiss it again. When I look at an EPC graph for my site from the time I started with AS to today, it looks normal... ups and downs. The magic happens when I make a graph with both clicks by day and epc. My EPC is right in synch with my Clicks per day... the higher my clicks go - the lower EPC goes... time and time again.

I don't know what to make of all this other than, it has led me to believe that there is a sliding scale at which we are paid, based on the number of CLICKS your site receives.

Why do I say based on Clicks?

- I noticed that if I improve my CTR, my EPC will adjust itself accordingly to make it so that I end up back where I was at before. I consider this an affect of Smart Pricing - and I ACCEPT IT AS SUCH.

The problem is have is when I simply increase my amount of traffic - FROM THE SAME SOURCE - and my CTR remains the same :::: and the EPC STILL falls off. The more clicks I get in a given day - the lower my EPC will be, almost guaranteed - this day, that day - every damn day.

Some things that I should note, as I know my post is disorganized... perhaps I can fill in some blanks.

+ I am not depleting my ad inventory. The ads are the same - ALL DAY, almost every day. There are several very large industry players in my niche that get the exact same adverts I get on my site... ver batim.

+ I believe that Smart Pricing is the reason I see my earnings "normalize" when I make a change to up the CTR - but not when traffic from the same source increases and EPC still drops.

+ SmartPricing is supposedly updated weekly... I notice that if Google SERPS shuffle, and my traffic goes through the roof like the flip of a switch - my EPC IMMEDIATELY falls off when the traffic starts coming in - and I mean immediately as in, within one hour the EPC is waaaaaaay lower than it was before the traffic flood started. It most certainly does not take them anywhere near a week to cut your EPC by 80% or more, it can happen over the course of 10 minutes - and then it sticks for about a week.

+ On the other end of the stick, if all of that traffic goes away as quick as it came, my site will continue to have a crap EPC (from high traffic levels) for a few more days. So the week theory checks out here.

So what is the deal really? Why do I always magically see this dramatic decrease in EPC (80-85% less than on days where my number of clicks remains relatively low is where it seems to bottom out)

I have switched 90% of my inventory to Overture, because I'm tired of games and theorizing about the internal workings of adsense. I want what is best for the advertisers too - so if there is something Smart Pricing doesn't like about more traffic from the same source - then it needs to be more transparent for me to do anything about it. I see no reason for such a sharp decrease in EPC... especially considering it can happen at the drop of a hat, and takes nearly a week to correct itself when traffic returns to "normal" in my experience.

At this point, even if Yahoo's CTR is garbage for me - I know what I am making per click, and it does not change based on weather or not I got 100 or 1000 clicks on a day to day basis.

Here is how I view adsense clicks in terms of value - please note these are made up numbers:

40 Clicks over the course of a day = .2 per click
40 Clicks over the course of an hour = .05 per click

100 Clicks over the course of a day = .15 per click
100 Clicks over the course of an hour = .03 per click

You get the point...

My final conclusion (more speculation) - is that AdSense discounts clicks based on the amount you are receiving over a given unit of time. I don't know if it is a static scale for all sites, or if they base it off of your site click history or even indivual page click history.

My big question is... Had my site STARTED OUT - with the level of traffic and clicks that I am getting right now - Would I be seeing such a low EPC? ... or is my low EPC the affect of having a staggeringly lower number of daily clicks on average over my entire time with adsense? All of the time I invest in finding out why my AdSense revenue does what is time wasted. I should be adding content RIGHT NOW - but the amount of monetary difference the EPC makes forces me to investigate.

I am going to give Google just enough traffic to keep them busy (just enough that my EPC stays high) - and Yahoo will get the rest until I can figure this mess out.

Any opinions, advice, or similar/opposite experiences welcome. I'm just trying to learn why this is repeatedly happening to me so that I can correct it.

[edited by: Brett_Tabke at 7:44 pm (utc) on Jan. 11, 2006]
[edit reason] fixed typo upon request [/edit]

 

aeiouy

5+ Year Member



 
Msg#: 11464 posted 5:41 pm on Jan 11, 2006 (gmt 0)

I think part of this relates directly to how adwords works. In some categories when new words are tried out then they pay a premium until they get a history going, and then the price they have to pay drops down.

I agree with others that there are so many factors at work here it would be difficult to accruately tell what is causing what to happen without knowing all the involved components.

david_uk

WebmasterWorld Senior Member 10+ Year Member



 
Msg#: 11464 posted 5:59 pm on Jan 11, 2006 (gmt 0)

What about those of us who are seeing the opposite "undeniable trend"?

Unless I'm seriously misreading the top post, as far as I can tell the OP is discussing a repeated pattern of increase in clicks due to an increase in traffic being invariably met by a decrease in EPC. When traffic and clicks subside again, there is a rise in EPC. The same trend reported by other posters here on many occasions since I joined the forum.

Are you telling us that all of your sudden increases in traffic are met by a increase in EPC, and a decrease in epc happens when traffic and clicks slips backwards again? That would be the opposite of the "Undeniable trend" being discussed here.

europeforvisitors



 
Msg#: 11464 posted 6:21 pm on Jan 11, 2006 (gmt 0)

Well, I'm not going to go back and calculate every rise and fall in traffic, clicks, eCPM, etc. over the last 3-1/2 years, so I'll just settle for saying that the "undeniable trend" reported by the original poster isn't happening for everybody, and it's always a mistake to think that one's own experience is a universal truth.

dethfire

5+ Year Member



 
Msg#: 11464 posted 6:37 pm on Jan 11, 2006 (gmt 0)

I have noticed this trend too. I am making almost 1/3 of what I was making a year ago with nearly the same amount of clicks :(

hunderdown

10+ Year Member



 
Msg#: 11464 posted 7:06 pm on Jan 11, 2006 (gmt 0)

Are you telling us that all of your sudden increases in traffic are met by a increase in EPC, and a decrease in epc happens when traffic and clicks slips backwards again? That would be the opposite of the "Undeniable trend" being discussed here.

As I said earlier, I haven't experienced the opposite in that sense, but a different kind of opposite--traffic/clicks and EPC don't relate to each other at all. And that suggests even more powerfully than the kind of opposite you describe that the phenomenon the OP describes is not universal.

europeforvisitors



 
Msg#: 11464 posted 7:41 pm on Jan 11, 2006 (gmt 0)

The original poster stated:

This brings me to this week. Google images has been sending me tens of thousands of referrals - destroying any prior records by mid-morning. My EPC continues to fall...

Maybe the source and type of traffic are the problem. Are visitors who have been searching for Google Images likely to convert for advertisers? I suspect not. (I do know that my eCPM on photo-gallery pages is abysmally low--it's a mere fraction of my eCPM from travel-planning articles and links pages, where readers are researching ways to spend their money.)

Jane_Doe

WebmasterWorld Senior Member jane_doe us a WebmasterWorld Top Contributor of All Time 10+ Year Member



 
Msg#: 11464 posted 7:57 pm on Jan 11, 2006 (gmt 0)

IMHO, you just have to get used to a different way of thinking when you have Adsense. With affiliate programs generally the more traffic you get the more money you make, but there is usually no downside to traffic increases. With Adsense this is not always true. Adsense sometimes requires a lot more testing and pattern analysis in order to increase your earnings, especially when you have relatively high traffic sites.

NoLimits

5+ Year Member



 
Msg#: 11464 posted 8:13 pm on Jan 11, 2006 (gmt 0)

I agree with you EFV, that images.google.com referrals probably don't carry a very high conversion rate.

The only thing that bothers me about what is currently happening - assuming that the conversions from my site are low with Google Images traffic is:

How can google know this on such an immediate notice, and adjust accordingly. How could they possibly have accurate conversion statistics from clicks that came through as short as 30 minutes ago. It just bothers me that the bottom can fall out of my EPC seemingly immediately - but it takes time to recover when the traffic subsides.

Does anyone have any statistics in regards to the conversion rate of images.google.com referrals?

Another thing that I'd like to mention that stings my nostrils with a bit of fishiness is:

If I cut the amount of AdSense ads I display down to say a 1:20 ratio (1 being AdSense and 20 being Yahoo) - I am showing essentially the same amount of AdSense ads as I was before waves of G-Images referrals started coming through. Once I do that, over the course of a couple days to a week, EPC returns to normal. So I feel like I am just sending them "more of the same" referral traffic, and receiving less.

The pages receiving this traffic were not intended to be big players in Google images... they are largely content based (500 to 1000 word articles with one image, 2 tops)

I don't doubt that the conversion ratio is crappy from G-images. I just wish I could confirm it for my site specifically so that I could justify spending time on working out a plan to improve conversions in one way or another... even if that meant making the ads less visible. I would prefer to have a solid EPC, knowing that both myself and the advertisers are getting a good deal, rather than droves of very cheap clicks... that don't make me - or the advertiser happy.

jonathanleger

10+ Year Member



 
Msg#: 11464 posted 8:29 pm on Jan 11, 2006 (gmt 0)

This trend makes perfect sense to me, and it also makes sense why it doesn't happen to everybody like (EFV).

Let's say there are 25 advertisers for your site's subject matter. When you have 1,000 clicks a day, the top 5 advertiser's budgets are large enough to support those clicks, and so you get top dollar for those clicks.

When your traffic surges to 10,000 clicks a day, however, their budgets run out early in the day, and so the other advertisers whose budgets are lower start to show their ads, and your eCPM goes down. Or, if the top advertisers see their budgets dwindling, perhaps they lower their CPC?

For other markets, such as EFV's market (which I believe is travel?) there is a huge amount of advertisers, and so the likelyhood of you exhausting their budgets is very, very small, and so people in those markets experience the opposite--more clicks do not result in a lower eCPM.

And as has been suggested, some forms of traffic convert very badly (i.e. image searches), which of course will lower your site's conversion rate for the advertiser, therefore resulting in a lower CPC for the advertiser and a lower eCPM for you.

I used to put some stock in the AdSense conspiracy theories myself, until I became an AdWords advertiser and started to realize how the other side of the equation works.

Now, Google is a large, stock-holder indebted company, and there's no doubt that they tweak things for their own bottom line's benefit, perhaps more now than they did before going public. But they are a business, and who runs a business for the charity of others?

I do think that Google needs to be more transparent. I don't think their "fear from silence" tactics are good for anyone in the long run, and I hope that competitors like Yahoo will start to force a more open approach. Canned email responses to serioues questions are not acceptable, and need to change.

But Google is not a dark mafia organization designed to take the AdSense publisher for a ride and throw him in the river, either.

21_blue

WebmasterWorld Senior Member 5+ Year Member



 
Msg#: 11464 posted 8:39 pm on Jan 11, 2006 (gmt 0)

NoLimits wrote:
If I cut the amount of AdSense ads I display... EPC returns to normal... I don't doubt that the conversion ratio is crappy from G-images. I just wish I could confirm it for my site specifically so that I could justify spending time on working out a plan to improve conversions in one way or another... even if that meant making the ads less visible.

Imho, I think you're on to something, here. But the lack of conversions data makes it impossible for us to work out what is the best publishing strategy.

If I move ads into a hot zone, CTR rises, EPC falls on that page. If I move ads out of the hot zone, CTR falls, EPC rises on that page. What happens to conversions? I don't know, but I guess I get a better conversion rate with the latter (partly because EPC overall on the whole site rises as well).

As an Adwords advertiser, I know that a well-worded advert can get a high CTR, with a low CPC. But it is interesting to think about the impact of such adverts on the publisher, if they don't convert well: it sucks a lot of traffic out of their website, for a low EPC, and depresses their SmartPrice.

In this context, how, as a publisher, do I maximise my income? Do I need to identify the "best" adverts with the "worst" landing pages and ban them? Do I need to move the ads in the highest CTR pages out of the hot zones?

What I ought to be doing is getting on and writing content. But a five minute tweak on a web page layout - to get fewer clicks but a higher EPC - can make more difference to the bottom line than 6 months of writing.

NoLimits

5+ Year Member



 
Msg#: 11464 posted 8:49 pm on Jan 11, 2006 (gmt 0)

What I ought to be doing is getting on and writing content. But a five minute tweak on a web page layout - to get fewer clicks but a higher EPC - can make more difference to the bottom line than 6 months of writing.

Thank you for reminding me why this is important to me. It's easy to just say, oh well - live and let live... Google knows what they're doing. (especially on a record earnings day)

But you are very right... one mundane detail could result in HUGE increases/decreases in overall income once you've got the traffic to play with. This is why I find myself constantly questioning the internal workings of the program. There's just too much money "on the table" to not take things like this to heart.

21_blue

WebmasterWorld Senior Member 5+ Year Member



 
Msg#: 11464 posted 8:56 pm on Jan 11, 2006 (gmt 0)

Google knows what they're doing

Well, as may be apparent from posts I've made elsewhere earlier today, I'm a bit worried that in some respects they don't know - they've created a system more complex than they realise.

Google's reassurances don't reassure me - what I've convinced by is hard data.

Jane_Doe

WebmasterWorld Senior Member jane_doe us a WebmasterWorld Top Contributor of All Time 10+ Year Member



 
Msg#: 11464 posted 12:18 am on Jan 12, 2006 (gmt 0)

I think there are discernible patterns to Adsense earnings programming, but it helps to have a variety of sites with different types of topics and traffic levels to get a feel for what is going on.

eyezshine

10+ Year Member



 
Msg#: 11464 posted 1:22 am on Jan 12, 2006 (gmt 0)

I think google is controlling the amount of money they are making by raising and lowering the money we make so that they will keep a steady flow of money coming in for their investors.

If their quarterly income is low, they lower our epc causing themselves to make their quota for the quarter to keep the investors happy. If they are making alot more than normal then they raise our epc so everything balances out in the end for their investors.

It's all about the investors but they also need to keep us happy too.

europeforvisitors



 
Msg#: 11464 posted 2:37 am on Jan 12, 2006 (gmt 0)

Eyezine, you obviously haven't read Google's quarterly earnings reports, which clearly state the percentage of revenues that are paid out to partners in the AdSense network.

hunderdown

10+ Year Member



 
Msg#: 11464 posted 2:51 am on Jan 12, 2006 (gmt 0)

Perhaps he has, but he believes that Google is not reporting the numbers truthfully. Personally, I don't think the financial guys at Google are so stupid as to do an Enron and file fraudulent statements with the SEC, but hey, it's a free country, everyone's entitled to their opinion.

Visi

10+ Year Member



 
Msg#: 11464 posted 3:10 am on Jan 12, 2006 (gmt 0)

As I have noted previously you need to read the financials carefully. There are a number of expenses included under the "so called" aquisition costs. Many of these categories increased by percentage greatly in the last reporting period. The percentage paid out as a direct cost to publishers probably has dropped since adsense inception, however no big conspiracy here. As google adsense has grown new hardware and support are required to substain it is all.

4string

10+ Year Member



 
Msg#: 11464 posted 3:57 am on Jan 12, 2006 (gmt 0)

Maybe a traffic surge trips some sort of click fraud protection scheme. Perhaps they don't view the traffic as trustworthy and are trying to protect advertisers. Or, they just like to spread the click fraud potential around to more advertisers. As a result drudging up many low paying ads.

When we look at our own sites are we sure we see the same ads others see? I'm doubtful.

danimal



 
Msg#: 11464 posted 4:03 am on Jan 12, 2006 (gmt 0)

>>>I've noticed that when viewing my pages from the g-images frame, most ads are entirely untargeted.<<<

i have not seen that problem so far, thankfully.

putting up adsense for google image results may not be the most effective thing to do, though... what i've seen is that the pages that google frames should be considered landing pages, only with a much smaller window to drive the traffic where you want it.

in my case, the google-framed site somehow got put in the sandbox, so i'm working on sending the google image traffic to a site where the adsense pays a whole lot more, via an attractive link visible in the google-framed page.

jhood

10+ Year Member



 
Msg#: 11464 posted 4:34 am on Jan 12, 2006 (gmt 0)

Everyone expects a quantity discount. If you buy 10,000 widgets, you expect to pay a lot less per widget than if you buy 10. The same is true of advertising units.

I have observed a roughly similar trend, though I would much rather have a lot of traffic and earn slightly less per unit than to have a little traffic and earn a lot per unit.

Google has to spread its paying customers over a huge, constantly growing and ever-changing landscape of sites offering inventory of wildly varying quality. It would not be surprising if it took a discount as unit volume climbs. That's life in the free market.

AS for YPN, I have found that the CPC and CPM are quite stable but consistently lower than I am able to get from Google, which pretty effectively ends that conversation. (I also find as an advertiser that Overture is much more expensive than Google, which pretty much ends that conversation too).

GoldenHammer

5+ Year Member



 
Msg#: 11464 posted 6:28 am on Jan 12, 2006 (gmt 0)

[....What about those of us who are seeing the opposite "undeniable trend"? ....]

My site experienced both trends, but more on the negative side (more traffic, but poor eCPM and earnings. Only a few on the positive side (more traffic, better eCPM and earning).

That seems to me there is no major difference on the traffic as over 85% are repeated traffic, I intended to believe that is the AS's Ad rotation algorthm. The increase of traffic has to "match" the AS rotation schedule on time .... :)

You are luck if your site always got postive trend, and you better hope it lasts.

david_uk

WebmasterWorld Senior Member 10+ Year Member



 
Msg#: 11464 posted 7:18 am on Jan 12, 2006 (gmt 0)

What about those of us who are seeing the opposite "undeniable trend"?

Are you telling us that all of your sudden increases in traffic are met by a increase in EPC, and a decrease in epc happens when traffic and clicks slips backwards again? That would be the opposite of the "Undeniable trend" being discussed here.

Well, I'm not going to go back and calculate every rise and fall in traffic, clicks, eCPM, etc. over the last 3-1/2 years, so I'll just settle for saying that the "undeniable trend" reported by the original poster isn't happening for everybody, and it's always a mistake to think that one's own experience is a universal truth.

Thanks for clearing that up. In summary you are in fact not seeing an opposite trend then.

Nobody has said it's a universal truth. Merely that it is a trend some of us have observed. In the case of the OP and myself the pattern is very clear. I'm looking at a graph right now. Others may not have seen the trend so clearly, others may not be analysing data that meticulously. Othes may not see the trend at all. However I don't see that anybody would see the opposite of the trend we are discussing as described above. Nobody is claiming it as a universal truth - merely a pattern some have observed more strongly than others.

In fact, it makes sense from a business point of view that Google would do this. Google isn't some sort of benevolent, indulgent great-uncle figure smiling down upon his favourte nieces & nephews - it's a global business whose primary goal is to provide a return for it's investors. Many of whom will have got shares on the basis of performance, and not have a clue what Google is / does, let alone care about it. They might think the name is cool though.

Google provides the best return to it's investors by maximising the amount of income from advertisers, and paying out the minimum to publishers. All Google's algorythms are geared towards this aim. Even though Google's payouts are good, it's clearly going to only be the minimum they can get away with, and will always be looking to maximise their profit. Nobody should be in denial that Google is a business, and doing what businesses do.

Therefore the trend noted is unlikely to be some accidental statistical blip, but part of Google's business plan. I think it's extremely unlikely that anybody can state that when the number of clicks rises, epc also invariably rises.

I'm not having dig, or whining about it. I merely make the point that it is what business does, and I accept that. After all, I don't do so badly out of my part in this business.

GoldenHammer

5+ Year Member



 
Msg#: 11464 posted 7:36 am on Jan 12, 2006 (gmt 0)

[.... If unhappy publishers leave AdSense for greener pastures, that could actually help Google by increasing the overall quality of referral traffic to advertisers (thereby encouraging higher bids). Why? Because publishers who are being "smartpriced to death" (to borrow a term that we've seen on this forum) probably aren't delivering clicks that convert for advertisers. .... ]

It is no doubt that Google optimized the AS algorthm to favorite the big advertisers and publishers. However, there is not such thing for a perfect configuration. The cost is the impact for some publishers in the opposite trend, they are not necessarily poor quality content sites, but just not perform well specificly under the existing norm.

I believe Google is smart enough not to make such "smartpriced to death" (great if they could make that for trunky websites), or they are at risk to make AS running into a monoply situation to favorite a specific type of websites or traffic models (that converts better at the moment). The AS algorthm just keep open and rolling.

21_blue

WebmasterWorld Senior Member 5+ Year Member



 
Msg#: 11464 posted 10:58 am on Jan 12, 2006 (gmt 0)

I'm not going to go back and calculate every rise and fall in traffic, clicks, eCPM, etc. over the last 3-1/2 years

You don't need to go to any great length to determine whether this trend applies to you. It should take about 3 minutes, as follows:

  1. In Adsense Advanced Reports, display aggregate data for "all time"

  2. Click on "CSV" and then "Open" to import the data into Excel

  3. In cell G2 type the following formula:
    +f2/c2

  4. Copy cell G2 down the rest of the G column to the bottom of the spreadsheet (stopping at the same row as the rest of the data)

  5. Delete the bottom row (ie the "totals")

  6. In a blank cell, say I3, type the following formula:
    =CORREL(c2:c9999,g2:g9999)

The result, in cell I3, will be a number between -1 and +1. If the number is negative then your trend is the same as the OP. If your number is positive then your trend is positive. (The only caveat is whether the result is statistically significant, but a 3rd party stats package is required to answer that question).

My result is negative: -.35. Ie my trend is the same as the OP's.

I would be interested to see others' results, because it is my guess that this trend is almost universal. Ie: I suspect that only be a tiny number of sites will have a statistically significant positive result.

GoldenHammer

5+ Year Member



 
Msg#: 11464 posted 11:44 am on Jan 12, 2006 (gmt 0)

Mine is -0.15 for a 30 months record.

NoLimits

5+ Year Member



 
Msg#: 11464 posted 1:15 pm on Jan 12, 2006 (gmt 0)

-0.11 Going back to Sept/04

hunderdown

10+ Year Member



 
Msg#: 11464 posted 3:39 pm on Jan 12, 2006 (gmt 0)

Mine is -0.12

What, exactly, does that number mean? Is 0 no correlation, and 1 in either direction a perfect correlation? Then -0.12 is a pretty weak correlation, isn't it?

In any case, 21_blue, thanks for that info. How could we modify this to run it for a subset of the data--all of us who have posted so far have data going back to before smart pricing was introduced. Smart pricing changed things. So I'm interested to see the trend over the past year.

hunderdown

10+ Year Member



 
Msg#: 11464 posted 3:46 pm on Jan 12, 2006 (gmt 0)

OK, I figured it out. Running the same formula for April 1, 2004 to date, I get -0.06, so a weaker confirmation.

But if I look at January 1, 2005, to date, I get +0.13.

21_blue, please interpret....

21_blue

WebmasterWorld Senior Member 5+ Year Member



 
Msg#: 11464 posted 5:15 pm on Jan 12, 2006 (gmt 0)

hunderdown wrote:
Mine is -0.12
What, exactly, does that number mean? Is 0 no correlation, and 1 in either direction a perfect correlation? Then -0.12 is a pretty weak correlation, isn't it?

That's almost right.

  • A number "near" zero means that the relationship between no of clicks and EPC is random
    OR
    that you don't have enough data to draw a statistically significant result

    (what is meant by "near" is defined in stats and could be a whole separate thread)

  • Plus one (and minus one) mean the relationship is a hard and fast rule - as the no of clicks go up, EPC also goes up (or down) in direct proportion

  • Anything not "near" zero is a trend; the bigger the number, the bigger the trend

OK, I figured it out. Running the same formula for April 1, 2004 to date, I get -0.06, so a weaker confirmation.
But if I look at January 1, 2005, to date, I get +0.13.

Alas, "interpreting" the data is really difficult, especially given the potential impact of SmartPrice. For example, one example of confusion might be:

  • Supposing that your number of clicks have gone up overall this year, it appears that EPC has also gone up at the same time. Is this a counter-example to NoLimits' observations? Could be.

  • But another explanation might be that your EPC has gone up this year independently of volume, because of SmartPrice. That is, there may be two trends at play, here: the NoLimits one, depressing EPC as your click volume increases; and the impact of SmartPrice, which has had an inflationary impact on your EPC. The latter has had a bigger effect, hence you get a positive correlation. But this is more of a coincidence of two factors than a simple, single trend.

To separate out these different 'factors' requires a more complicated statistical analysis, called 'factor analysis'. This separates the confusing correlations, to some extent.

Unfortunately, Excel doesn't have a function for factor analysis. However, the last time I did a factor analysis on my stats it suggested that my EPC goes down in response to three things:

  1. as number of clicks increase
  2. as page impressions increase (puzzling, though this may be related to the CPM Ads problem)
  3. over time generally (also puzzling)

As a final aside, the factor analysis it also suggested that the biggest factor by far in increasing overall earnings was number of visitors.

europeforvisitors



 
Msg#: 11464 posted 5:29 pm on Jan 12, 2006 (gmt 0)

As a final aside, the factor analysis it also suggested that the biggest factor by far in increasing overall earnings was number of visitors.

That sounds reasonable, if only because:

1) "Ad blindness" is less likely to be a problem, since new visitors won't have seen the same ads day after day.

2) On many (not all) sites, new visitors who arrive through search will be researching topics that involve possible purchases.

3) On scraper and other made-for-AdSense sites, users aren't likely to stick around or come back, so increasing the number of visitors is the only way to boost income.

Side note to David_UK: I disagree that the original post wasn't implying a "universal trend." The subtitle of this thread is "The greater your number of clicks, the lower your EPC goes." Unless the "you" referred to some unnamed individual, the subtitle suggested that the trend was general and broad-based.

hunderdown

10+ Year Member



 
Msg#: 11464 posted 5:38 pm on Jan 12, 2006 (gmt 0)

21_blue, thanks again. Your language is more precise but that was pretty much my understanding of it.

The interesting thing is that since January 1, 2005, I made some changes to my site (primarily by adding an AdLinks box to my home page and redesigning the site setup, which changes how people flow through the site), that have greatly increased clicks, which now average nearly double what they did a year ago. So it seems to be possible to do things that will counter the + clicks = - EPC correlation....

Your point about more visitors reminds me that I have some work to do in that area.

This 70 message thread spans 3 pages: < < 70 ( 1 [2] 3 > >
Global Options:
 top home search open messages active posts  
 

Home / Forums Index / Google / Google AdSense
rss feed

All trademarks and copyrights held by respective owners. Member comments are owned by the poster.
Home ¦ Free Tools ¦ Terms of Service ¦ Privacy Policy ¦ Report Problem ¦ About ¦ Library ¦ Newsletter
WebmasterWorld is a Developer Shed Community owned by Jim Boykin.
© Webmaster World 1996-2014 all rights reserved