Forum Moderators: martinibuster
I'm wondering how widespread this trend is. Any other experiences with a large drop in Adsense revenue in the last few months? Whether your answer is yes or no, it would be helpful to have a ballpark estimate of site traffic.
Frankly, I'm sick of the "smart pricing" response to explain revenue fluctuations. No rational approach would decide overnight that advertisements are worth 25-50% of what they were yesterday. If the trend is true, then I suspect something other than smart pricing is afoot.
If the trend is true, then I suspect something other than smart pricing is afoot.
I can't answer you question about drops for ups club. If there are drops, then it fits with what I've been anticipating since about April, 2006, and that is that there "should" be an advertiser behavior shakeout in late 2006 and in 2007. As with earlier booms/shakeouts in internet ads, there's a big burst of paying more for ads, followed by a decline over time as advertisers realize that they can't get sufficient ROI to justify the costs in a competitive bidding marketplace.
So, my guess is that if drops are occuring, it is at least in part due to advertisers dropping their per click rates, while at the same time keeping their overall budgets stable or higher for 2007.
Higher volumes at lower per click costs.
Starting January 2nd, my eCPM increased for no reason I could discern. This increase was enough to propel me toward UPS club territory. It held at the new level for awhile, then just this week dropped down to below my previous average (before the increase). Again, I don't know any reason for this. Smartpricing explains neither change.
The way I see it, I can't complain too much about the low periods if I enjoy occassional high periods. I just wish I understood what causes the fluctuations. Maybe something to do with the phases of the moon..?
I interpret this to be standard smart pricing at work, though. Schools and universities provide a lot of my traffic, and now that they're back in session I'll be surprised if the numbers don't gradually drop back down to the average.
My traffic has been growing for several years, last year AS incoming did not grow any more. Few weeks ago AS earnings started to drop to 60~ 40% down.
Two days ago I made an extensive cleaning process and reduced my ads units in 50%. Basically I leaved only one ad unit per page (the best ctr). Surprisingly after that, my cpc decreased, my ctr decreased (?), and of course my eCPM decreased.
What is the response time of the smart algo?
Am I wrong in expecting a better cpc with less ad units?
I hope this may help, and thank you in advance for any suggestion.
Am I wrong in expecting a better cpc with less ad units?
So, Google keeps things fluctuating enough that the publishers can't get a handle on what is normal or fair and what isn't.
Google doesn't keep things fluctuating. The market keeps things fluctuating. AdWords/AdSense is an auction-based system. It's like the stock market or a mutual fund. (Dang! My Widgetco shares went down 5% yesterday. I wonder if NASDAQ is skimming share value, or if the mutual-fund manager is keeping a bigger cut?)
Also:
- You're assuming that all publishers are seeing big day-to-day or week-to-week fluctuations in EPC. That isn't the case.
- If you've ever worked with a rep firm, you'll know that display-ad prices vary quite a bit, too--often to a greater degree than publishers' EPCs from AdSense do. That doesn't mean the display-ad sales rep is pocketing a bigger commission on the Whatsitco banner flight when the CPM is low or a smaller commission on the Gadget Corp. skyscraper ads when the CPM is high; it just means that, in a free market, pricing is largely controlled by supply and demand.
ADDENDUM: On my own site, I see far more day-to-day variation in revenue from affiliate programs than in AdSense revenues. (And I've got a pretty large statistical sample, since my affiliate programs generate more revenue than my AdSense ads do.) AdSense publishers who are uncomfortable with fluctations might want to find another line of business--such as a salaried day job--because in most kinds of publishing, revenues go up and down with the season, with changing economic conditions, etc. (That's why magazines often seem thicker or thinner, depending on the time of year, and it's why a typical U.S. metropolitan newspaper is a lot thicker on Sundays than on a weekday.)
[edited by: europeforvisitors at 3:47 pm (utc) on Jan. 19, 2007]
Been there, done that, not fun.
FWIW, I'm seeing an increase over the last few months, not a decrease.
[edited by: incrediBILL at 3:44 pm (utc) on Jan. 19, 2007]
Possible causes:
"smart pricing" - does smart pricing kick in once a year and adjust by 60%? If yes, that's dumb pricing. I've never bought into this concept.
"loss of a big advertiser" - doubtful, since there's rarely just one ad/company advertising so several had to have similar CPC bids.
"change in adserving algo" - most likely. This would affect only certain sites that get "recategorized" by the algo, so they get slightly different ads that have a potentially very different set of bids.
I don't think Google is capricious or manipulative - I think they genuinely work to give the best ROI for their advertisers. The problem is that large-scale changes that happen without warning can be problematic. I don't think getting a day job is the answer - I think mitigating risk is. Erratic performance from Adsense pushes us to diversify revenue sources, but that costs us money too.
Would it be helpful for G to put a sliding average time window on price changes? This would slow down both revenue drops and increases, but it would make things more predictable (except you wouldn't know where the bottom or top would be). Oh I don't know...
Did Google change their formula for payment to publishers?
Some publishers have repeated increases in EPC, eCPM, or total earnings. So, if there was a change, it's helping some while hurting others.
Of course, you need to have some AdSense ad tracking software installed to see that the fluctuations are in the same direction across the board.
If you have several websites on different topics, you will see Google ramp the payout to your publisher ID up and down without respect to the subject matter (thus, not due to specific topic advertiser budget fluctuations). This only makes sense if Google is ramping the publisher share up and down from day to day or week to week
That's not the case with us, or rather it's kinda the case, and mostly not. Our sites move independently from day to day in terms of EPC/ECPM, and the variance within a site from day to day is big - much bigger than a year or two ago.
BUT, there's also an overall effect that seems to work across sites and content also.
"loss of a big advertiser" - doubtful, since there's rarely just one ad/company advertising so several had to have similar CPC bids.
it makes it clear that you don't understand auctions.
When you have one big player that will pay whatever it takes to be #1, someone else that wants to be #1 is willing to pay X and a third party that is willing to pay even less at .5X.
While the big player is in there, the bidding will go to X+1, when that big player drops out, the bidding will only go to .5X+1. Just short of a 50% drop.
I see this at real auctions all the time. Two startups show up at an auction looking to furnish their new offices and office chairs that would go for $20 each any other day, get sold out at $80+ each.
Most likely what you are seeing is a combination of factors, so don't throw out one before you understand how powerful that factor can actually be.
While the big player is in there, the bidding will go to X+1, when that big player drops out, the bidding will only go to .5X+1. Just short of a 50% drop.
Let's say you're right that a single high-bid advertiser dropped out, leaving the entire field of competing ads at a lower eCPM. Would it make sense to run a site-targeted Adwords campaign on your own site at a high CPM to bring up the rest of the bids?
Seems that you would get the publisher's cut back on the ads you run. So if analysts are right, you'd be paying a net 30% of your bid to raise the eCPM for at least one other advertisers on your site by the original 60%. Is that against Adsense or Adwords TOS?
The math works to a profit if the original adsense drop was big enough.
Would it make sense to run a site-targeted Adwords campaign on your own site at a high CPM to bring up the rest of the bids?
No, because you'd simply get your own higher-priced site-targeted CPM ads instead of the lower-earning CPC ads (which would be displayed on other publishers' sites).
Since October 06, only 3 months ago, I have seen a 50% drop in my revenue. Over the past year I have seen my numbers drop by 75%. In all seriousness my December 2005 was 4x the revenue of December 2006. This might seem like a long time span and give ‘smart pricing’ an easy justification, but keep in mind that my December 2003, 2004, 2005 were all within $2,000 of each other. At my numbers that $2,000 swing from month to month is pretty typical, but always staying on their side of my average in the high five digits.
This new decline has been occurring as Page Impressions and CTR go up. The explanation I get from my rep is that spending is down and it is affecting my vertical.
Does Google think that this is a legitimate explanation? When CPC have been going up especially during the holiday months, as I have seen for the past 3 years.
Just to put my pennies worth in here. Whilst this is the case in YPN & MSN - Adwords does not work this way. It does not run on a standard auction system. You can be placed at number one position on a highly competitive keyword - with some people bidding bidding tens of dollars for keywords, but only have to pay cents. This is what Adwords call their QS system - which means the amount you bid, and your ad placement is based on a number of factors including landing page quality, quality of ad text etc etc.
Re UPS fluctuations. Adsense is in permanent flux. The trick is to take the rough with the smooth, continue to write quality content, and make sure all your eggs are not in one basket.
For myself (in UPS) I have noticed a significant rise in traffic, but at the same time my EPC has declined a bit. The overall result is that monthly earnings have increased about 30%. This is completely normal behaviour based on two years worth of data. The EPC decline is partly accountable - for me - in the exchange rate fluctuations.
As you have read here - nearly everyone has ups and downs, it is necessary to look at the larger picture - don't look at one months figures look at quarters, or annual comparisons. Smartpricing - who knows, this has been discussed in detail elsewhere - and all anyone here is doing is groping in the dark for a pattern or answer, when better time could be spent on improving the site, and if adsense is not producing the right revenue for you, then investigating some time and effort into finding other means of monetising your sites. :)
Maybe it's just air. OR maybe G$$gle needs to pay for new datacenters and offset the cost of banning MFA networks recently...
Whilst this is the case in YPN & MSN - Adwords does not work this way.
That IS the way the auction part works, but as you say, the other factors play a part in the positioning. You basically get a discount for your high CTR, but it is still an auction.
As you say, if you have an ad that gets a high CTR, you can get away with bidding less and still get that high position. But if the big players have bid up the "value" of that top spot, you have to increase your bid to stay in the competition for that top spot.
If a high bidder drops out, you can easily see your EPC drop 50%, as everyone else no longer has to compete with them, including those with high CTR.
and offset the cost of banning MFA networks recently
Or maybe the UPS club sites were making all their money off the MFAs supporting the ad prices.
As long as we can see that there are a big number of webmasters on WebmasterWorld affected by this drop, we can call it a statistically significant event. Imagine how many webmasters out there don't watch numbers or don't post here - plenty. So I would suggest dismiss talks like "you were smart-priced", "improve your ads" or "improve your site" as insignificant and even counterproductive to understanding of the mechanisms of the drop.
So far I can only see 4 possible conclusions given the above as a premise:
1. There are websites who greatly benefited from this tweak - by "greatly" I mean who saw a 50% jump.
If this is the case, can someone who saw a 50% earnings jump in November/December 2006 please stand up now!
2. A big number of advertisers dropped out in that particular month.
3. There's new mathematics in the Adsense algo
4. Google played with algo so that they can squeeze mo money from the ads
My money is on #4.
[edit]added #3,4[/edit]
As long as we can see that there are a big number of webmasters on WebmasterWorld affected by this drop, we can call it a statistically significant event.
No, we can't. In fact, it's so anecdotal we can't really generalize ANYTHING. We don't have anything close to a random sample, which is what we'd require for statistical significance to generalize.
My money is on #4.
In that case, you can come back to crow or apologize when Google releases its 4Q 2006 numbers at the end of January. :-)
