Google Instant's Impact
I’m writing this blog on Halloween, so first off, Happy Halloween to all of you!
Google Instant has already changed the way people search the Internet. I read a detailed article on the extensive research that Google did to design the user interface, so obviously many people must like it. I personally do not. I like doing long tail searches and do not want Google assuming they know what I’m looking for as I type my search.
But, this blog is NOT about the end user side of Google Instant. It’s about the impact it is is having on eCommerce sites that rely heavily on traffic from organic and Pay-per-Click advertising on Google to acquire new customers. That would be most of us the last time I looked.....
My question is - have we been tricked or is there really a treat there someplace for our businesses?
Disclaimer: All businesses will likely have different impacts from Google Instant. This blog details a few impacts on our business. If you take anything away from this, check out what’s happening in your own business and make necessary adjustments.
Initial Results Promising
I was asked by Practical eCommerce shortly after Google Instant launched back in mid-September whether our referrals from Google were impacted by that changeover back on September 8th. At that time, the answer was yes, traffic had actually increased a bit from organic search on our two largest websites.
In the same response, I expressed concern because the traffic was coming from different keywords and it was unclear whether that was going to be a positive or negative impact to our overall sales.
2 Months Later - More of a Trick than a Treat
Before I get into details, let me say that my observations here are not from extensive personal analysis of our organic and PPC analytics data. Rather, they are based summary data from our various analytics and actual sale results. But, I’ve seen enough data to decide to move a fair portion of our PPC advertising to other media venues for the time being.
Historically, our sales in July-August are always less than the September-October period. This year, referral traffic from organic search from July-August is almost identical to the period from September-October. Our referral traffic should have been up significantly in the latter time period, so that’s not good. Our conversion rate is ever so slightly higher in the September-October period, but not as high as past years.
Pay-per-Click (PPC) traffic is a different story. The very obvious thing we see is that our cost per acquisition (CPA) has increased dramatically. In pretty much every campaign, every ad group and for every keyword, we see the same trend. Our conversion rates are down, and our cost per click is going up.
PPC traffic is up, by our own choice, by about 36% as we launched a significant new campaign just prior to the launch of Google Instant. Admittedly, this skewed our initial results and makes it a bit more difficult to compare periods, but we've sorted through that initial transition.
Total revenue from all our Google PPC campaigns is completely flat. Overall, our revenue is up by a normal amount in Sept-Oct from July-Aug, but our Google PPC campaigns are showing no growth. Average order sizes are about the same, but our CPA has gone up by about 65% since August. I expect some increases in CPA and CPC with any new campaign, but not that much, and not 2 months later after several rounds of adjustments. Not to mention, all our other campaigns have been affected in the same way.
The net of all this is that current cost per acquisition is way too high for us to pay going forward,
Why is this Happening?
It’s fairly obvious that our long tail keywords that have traditionally produced conversions are not being clicked on in our PPC campaigns as frequently. Same thing on organic search results. My guess is that broad terms are back “in”, at a much higher cost to us. I've seen some buzz about that from other sources as well.
Since Google is “presenting” mostly broad terms in the Google instant results and ads for those terms are being shown instantly and changing constantly, they are most likely getting the majority of the clicks from consumers. We certainly can increase traffic by focusing on broad terms by being in the 1-3 position, but our bounce rate goes off the charts and ends up being a waste.
For us, broad terms convert poorly and they are expensive. For many of you, especially if you are promoting brand goods, broad terms might be a good thing. But over time if if you compete against Ebay, Amazon, Buy.com, Overstock.com and the Brands themselves for those terms, you can bet your cost of advertising will increase.
What to Do About It
We’ve added dozens of negative keywords to our broad termed campaigns in the last few weeks. We’ve stopped some of the bleeding, but we are having a very hard time identifying new keywords that are delivering a real return. We’ll continue to cut off ads and keywords that are unproductive and bid up things that are working. Basic blocking and tackling.
In the short term, we're moving some of our Google ad budget to Microsoft in hopes that we can boost our Yahoo and Bing traffic and conversions to offset the Google issues. If we see improvements in Google PPC, we’ll start to reinvest there again. If we can draw higher conversions and enough traffic, we'll keep investing in Microsoft.
I suggest all of you take a really close look at what’s converting and make appropriate adjustments. It’s not easy, there is a lot of data to sift through and a fairly short history at this point to make decisions from. But, we all know that runaway keywords and campaigns are a huge cash burn.
Conclusion
Obviously, Google knew that Google Instant was going to increase their ad revenue. Just as they did extensive research on the consumer side, you can bet they did even more research analyzing the revenue impact of showing broader terms more frequently.
I think we’ve all been tricked - no treat in sight this Halloween!
1 Comment
researchinc says:
I have been going through similar process but without expending a lot of energy at this point. Even surface digging indicates long tail clicks and even impressions are down. Conversion seems to be suffering as well and it represents no trivial frustration. You really pointed out the weaknesses in their new model from a customer's perspective. I originally was not concerned at all (did not perceive the potential problems) with this new function but now understand how (and I think why) it is impacting our PPC returns. We are unhappy to say the least but grateful for your article!
Connect with us