RadarFarms Press

We are about to launch our new advertising offering – 77Radars - and we are running Case Study to understanding how this offering is stacking up against two pillars of Internet Marketing – Natural Search (a.k.a Search Engine Optimization or SEO) and Paid Search (a.k.a Pay-per-Click or PPC).

Therefore, we decided awhile ago in mid-April, 2007 to set up a PPC campaign for our two radars dedicated to niche subjects – Tuvan Stamps and Vintage Massandra Wines.

We decided to use services of an absolute champion of all PPC services – celebrated Google AdWords. Don’t get us wrong – we all know and admit that the service is excellent, or otherwise Google would never be the Google as we know it. However,  our impressions from the set-up were mixed.

The AdWords user interface is not bad but for an individual who never done it before, it is sort of too complicated.

Initially, after the account setup, one should set up a max Cost per Click and daily budget. We set up $0.12 CAD as a max cost-per-click (CPC) and $0.42 CAD as max daily budget. It would match $11 of targeted monthly cost for sponsors of 77Radars.

Google scanned the Massandra radar’s Recent Items page and came up with some keywords. I added “Massandra” and added the keywords I felt as most appropriate. Then I used “View Traffic Estimator” tool by Google. The tool allows estimating average CPC, Ad Position, daily Clicks and Cost.

For Massandra, the tool estimated that the average cost would be $0.09 CAD and 0-1 clicks a day.

After we set up the Massandra campaign, we ran Google AdWords External Keyword Tool and the keywords it proposed were different from the internal tool.

Then, we checked the campaign status and turned out that 3 keywords related to Sotheby’s were out of reach because they started at $0.23-0.46 CPC that wasn’t indicated by the internal tool initially.

The bottom line – Google didn’t have any bidder for “Massandra” keyword yet but us. We were going to add some keywords later but the whole process was not for a mere mortal – it is for sure.

The same process was a bit less complicated for Tuvan Stamps which was set up by me for the same CPC and daily budget as the Massandra radar. Google has one bidder for “Tuvan Stamps” keyword, which is ranked as Pos 1.

We added “rare stamps” keyword, where we were OK from the CPC/budget’s standpoint but, according to the Traffic Estimator, could have positioned as 4-6. When we checked Sponsored Links for “rare stamps“, we were not there.

The Traffic Estimator informed us that the average CPC would be $0.06-0.12 and number of daily clicks would be 0-1, however, for $3.35 max CPC and daily $10 budget, they will bump me up to #1 position for the keywords picked by us.

We used the External Keyword tool for the Tuvan keywords, and it wasn’t very different from the internal tool.

Again, we initially intended to fine tune our campaigns, but then we decided to leave them alone and probably this way we just mimicked behavior of an average “Advertiser Joe” who will not fix it if it ain’t broken.

We haven’t noticed too much traffic brought to us through AdWords but then on June 30 our campaign for the Massandra radar ended. We don’t remember that we set up this date but nevertheless decided not resume it because for 2.5 months since the campaign was launched, we just received one visit via this source.

We checked the campaign and had quite a few questions though:

All our keywords had “Inactive for Search” status.

Google defines this status as “A keyword is marked inactive for search if it doesn’t have a high enough Quality Score and cost-per-click (CPC) bid to trigger ads on Google or the search network. This means your keyword’s CPC bid doesn’t meet the quality-based minimum bid.”

Three out of ten keywords selected for the campaign had OK Quality Score so it must have been a low bid. Indeed, Google asked for the OK keywords to increase minimum bid from our low-budget $0.12 to $0.35 – not too shabby. That’s how the price went up. It was surprising, because “crimea ukraine” still doesn’t have any bidders while “Massandra” keyword just has one – eBay. We cannot help but make a conclusion that Google decided that it will be better off not serving any ads for the “crimea ukraine” at all, rather than having a low-paying advertiser. The same conclusion could be applied for “Massandra” where apparently eBay shelled out required $0.35 per click and mighty Google decided that it doesn’t want the lower bidders.

Let’s fast-track now and check our more successful AdWords campaign for Tuvan Stamps:

For 6.5 months since we launched the campaign, we received 302 clicks – not bad indeed. However, as of today all five keywords we selected are inactive for search. It is the same pattern as we saw with the Massandra campaign. Keywords with high Average Position and OK Quality Score are inactivated, and Google demands more monies for a minimum bid. The increase is not as dramatic as it was with Massandra but still the CPC went up from $0.12 to $0.18 and “Tuvan Stamps” keyword has just one bidder while “Tuva Stamps” keyword has two bidders.

Anyway, now we need to make a decision whether we will continue the campaign or let it end.

We also noticed that even when the above keywords were active, our ad had not been always served for the reasons unknown to us.

New metric by Google “Impression Share” helped us to understand why. Google defines it as: Impression share is a new metric that represents the percentage of times your ads were shown (i.e. your accrued impressions) out of the total number of page impressions (i.e. pages where your ad appeared or could have appeared) in the market you were targeting.

It means that Google is not always serving ads even if it could. Here we need to give a credit to Google because via their advanced reporting one could understand why Google is passing on an ad. They allow the AdWords user to add couple of sub-metrics to the Campaign report and see whether the Impression Share (IS) is lost due to poor Rank or low Budget (see our Campaign Reports below):

That’s how Google defines these values:

The ‘Lost IS’ columns reflects a portion of the total page impressions you were eligible for that were lost due to budget or rank reasons. To illustrate,

Your impression share + Lost IS (Budget) + Lost IS (Rank) = 100%.

Please note that we do not report lost impression share measures below 10% because the values may not be meaningful based on the volume of data available. These values are expressed as ‘N/A’ in the columns.

• Lost IS (Budget)
  You are losing impression share due to an insufficient budget.
• Lost IS (Rank)
  A poor Ad Rank (cost-per-click bid x Quality Score) is reducing your impression share. You may want to try adjusting your keyword match type to achieve a higher impression share.

Good news – we didn’t lose any IS due to low monthly budget, bad news – we lost 86% chances to be served for our Tuvan Stamps keywords apparently due to low cost-per-click bid.

We were so “impressed” by Google Impression Share (no pun intended) that decided to start monitoring the subject and created AdWords Impression Share radar.

Hopefully, our little AdWords saga may help niche online vendors or little guys like us to make an intelligent decision on how to spend scarce advertising dollars.

All and all, Google AdWords is created to maximize the Google profits (nothing is wrong with it) and if your niche is too narrow and not quite popular in terms of web traffic, Google trying to offset its overhead expenses to serve your online ad with increasing CPC even if there are not any bidders or just few of them.

We also learnt that in order to run the successful campaign, you should monitor it closely yourself or, if you are not able to do it for whatever reason, outsource to AdWords campaign managers/consultants. In the latter case, hiring the AdWords pundits could significantly increase advertising costs – see our AdWords Campaign Management radar.