Search Engine Monkeying

I haven’t posted in a while, I’ve been busy working and not reading RSS feeds, but two separate bits of news caught my attention and I was wondering if anyone else thought there was something more to it.

A couple weeks ago, Google confirmed that they were including load times into the quality score algorithm.  SearchEngineLand asked Google a series of questions about the new load time variable which Google responded too.  The question on everyones mind was question 4, “How many seconds is considered a bad page load time?”  Google answered “Our goal is to encourage advertisers with landing pages that are excessively slow to speed up their sites. It’s not possible to provide an exact number of seconds above which we would consider a site to load “slowly,” since we adjust for regional and country differences in average load times.”

Today. I was reading SearchEngineLand and they were talking about making your site “geospecific” and how Google determines where your site is located.  How much are the two projects related do you think?

Popularity: unranked [?]

You may have already noticed my new Google Talk badge in the sidebar.  Pretty cool considering all you have to do is copy an iframe and paste it into your code.

Unfortunately, now I’m going to have to turn off my invisibility on Google Talk.  It was fun while it lasted.

Popularity: unranked [?]

On Feb 28th Google AdWords is preparing to bring their new “automatic fast matching” out of beta. The whole idea sounds like a bad upgrade for users and a way for Google to make even more money.  There are three things that really bug me.

• Related terms do not necessarily meant that the search term is in the consumer buying funnel. SEO Fast Start mentions the term Addidas being related to the term shoe being related to the term slipper. Not even close. It would be a disaster if this happened.
• Automatic Fast Match spends up to your daily budget. I don’t know about most people, but my daily budget is more of a safety net then an actual amount I want to spend. The idea of junk terms raising my spend is almost inconceivable.
• From what I’ve heard, everyone’s account is automatically going to be opted-in. What???

maybe I’m being too harsh and I’ll feel better about this additional feature after I see the reports that are included. Right now, I’m not very anxious to give it a try.

Popularity: 100% [?]

The Microsoft AdCenter blog published this great illustration of the customer buying funnel.  I have a hard time explaining this to businesses that I work with and illustrating the value of targeting different consumers in different steps of the funnel.

What makes this more complex is if you look at the relationships that exist in the B2B world.  The fact that the consumers who may visit your site and design products are not the actual people who purchase the product for the company make this much more difficult.

At SES last year they discussed this phenomenon in one of the seminars on Landing Page testing.  The business had their website configure more for executive and purchasers but when they targeted the engineers who designed and used their products sales went up gradually.

Additionally, the customer purchase funnel is a major component in contextual ad marketing and pricing.  Perry Marshall suggests that contextual keywords on Google should be only 1/2 the bid of the same words in the search network.  I altered the bids on one of my campaigns and found this to be absolutely true and saved a lot of money.

Popularity: 91% [?]

Look. Look.  The Google Checkout tag does get attention.

This ad must get twice the attention.

Popularity: 88% [?]

These are a few of my favorite things:

• pg 76 - Relevance between the Headline and the Display URL.
• pg 90 - Long Tail keywords don’t work (indirectly related terms).
• pg 101 - Local ads help organic search results on Google.com.

There is also some interesting info on Alternative Search Engines (2nd
Tier) but the comments are totally contradicting. Did anyone else see
anything else particularly interesting? Do you think some of the
conclusions drawn from any of the experiments is incorrect?

Popularity: 87% [?]

Nice article that sums up how to look for bidding software.

1. Don’t even think about building it yourself. I speak from experience here. Building bid management software requires a full-time team, ongoing maintenance, and a lot of trial and error. It will take you at least a year to build a basic version, and at least two to three engineers to maintain and iterate it after that. And it won’t be as good as the software currently available on the market.
2. Assess your expertise and what you really need. Assuming you listened to my first tip, you next step is to understand how you are going to use the software. First, let’s talk about your level of expertise. If you are an expert, you may want to let the bid management software run your tail terms (the 98% of keywords that make up 2% of your revenue) and focus on optimizing the head yourself. If you aren’t an expert, you probably need software that can manage everything for you, with a very simple interface, and possibly the option of full-service bid management combined with the software. Either way, you need to know exactly what you want before you start talking to software providers. Otherwise, you might end up paying for a Ferrari when all you really needed was a station wagon.
3. Understand implementation and de-implementation effort and impact. A lot of bid management software only works if you install a snippet of code on your Web site and if you allow the bid management company to change your URLs on the search engines. This can require significant effort by your internal tech team and changing your URLs in your search campaigns can result in a loss of keyword history (i.e., you will need to pay more to get the same position). Moreover, you need to understand what happens if you end your relationship with the company - will they change your URLs back, or are you stuck with their tracking for the rest of your life?
4. Always do a trial first. I’ve seen some really great PowerPoint presentations from bid management companies. It turns out its easier to make a good PowerPoint than it is to make a good bid management software. Never sign up for anything until you have taken it for a test drive for at least one month and if possible three or four months.
5. Set benchmarks for initial and ongoing success. Before you start any trial, understand the status quo of your campaigns. What’s your current revenue? Profit? Margin? Tell the bid management company your actual metrics and tell them what you expect them to hit for them to win your business. Make sure you factor in the cost of their services. For example, if a bid management company wants to charge you 5% of your spend, and you currently have a 10% margin on your spend, you should demand that they at least bring you 15% margin (and probably higher). By the way, most bid management companies will thank you for this - it gives them something tangible to shoot for!
6. Look for hidden fees. Does the contract include API costs, or do you have to pay these? Is there a charge for consulting and implementation? Is there a minimum monthly bill? Read your contract carefully and ask a lawyer for help if you are at all confused.
7. Ask for performance pricing. I know my co-panelist Kevin Lee is going to kill me for saying this, but don’t be afraid to ask your bid management company to put some skin in the game. If a company’s bid management software is a good as they say it is, offer them 50% of the incremental profit they make you to prove it! More realistically, perhaps ask them to take a slightly lower percentage of spend in return for a performance bonus if they achieve certain goals (see Kevin, I’m not as unreasonable as I first seem!)
8. Get a short contract. If possible, try to get a month-to-month contract (though this will be hard to do). If you can’t make this happen, a six month contract is usual very doable.
9. Be hesitant about handing over your head keywords. For the 50 to 100 keywords that drive most of your revenue, I usually recommend good old human management. Why? Well I believe that a good search analyst just gets an almost intuitive feel for how to grow top keywords, something that computers just can’t do. And managing your top keywords in-house can save you a lot on bid management fees, especially if less than 50 keywords make up 20-30% of your ad spend.
10. Keep testing new competitors. The bid management world is ever-changing. I see new and exciting bid management companies popping up regularly. Always keep a campaign or two available for the next great thing.

10 Tips for Choosing Bid Management Software | Search Marketing Standard Blog

Blogged with Flock

Tags: BidSoftware

Popularity: 79% [?]

• Create a manageable, targeted keyword list.
  Advertisers have found most success on the content network with ad groups of around 15 to 30 keywords.

• Use tightly themed ad groups.
  For contextual targeting, we look for pages that match most of the keywords in your ad group. For example, if your ad group has a number of keywords about lilies and tulips, we try to find pages about these two topics together. If you have an ad group with diverse keywords on different themes, it may decrease the number of pages on which your ad is likely to appear. When picking keywords, imagine what keywords would likely appear on the pages that you are trying to target, and create tightly themed ad groups around those keywords.

• Use duplicate keywords for appropriate ad groups.
  To continue the previous example, let’s say you were creating a campaign for flowers and had ad groups for lilies, roses and tulips. Unlike search, we would recommend that the general keyword flowers be included in all three ad groups to help establish a floral theme.

• Use ad group level URLs instead of keyword level URLs.
  Because no one particular keyword is used to trigger your ads on the content network, keyword level URLs are not relevant. We recommend using ad group level URLs instead.

• Measure content performance at the ad group level.
  We’ve found that measuring your performance on the content network at the ad group level offers a better gauge of what strategies work best.

• Build a comprehensive negative keyword list.
  The more negative keywords you include on a particular topic, the less likely your ad is to appear on pages that match that topic. If a page is predominately about your negative keywords, while partially about your positive keywords, our system is not likely to show you on that page. If a page is principally about your positive keywords, but mentions a few negative keywords, then your ad may still appear on this page. We recommend that you include multiple negative keywords on topics you would like to avoid. If you sold camera film and wanted to reduce the likelihood of your ad showing on movie-related pages, you should include multiple negative keywords like -movie, -movies, as well as synonyms like -cinema, -cinemas.

Inside AdWords: Google Content Network Tips: Part 3 – Optimizing your keywords for the content network

Blogged with Flock

Tags: GoogleAdWords Content

Popularity: 79% [?]

Google has upgraded its Ads Diagnostic Tool, a tool that allows AdWords advertisers to determine why their ads may not be showing on a given SERP. Where it previously listed just one reason why an ad might not appear, it now lists multiple reasons, and multiple issues that should be fixed.

The tool is accessible within an AdWords account, either via a magnifying glass icon next to an individual keyword in an ad group or in the account’s ‘Tools’ page.

Google Improves AdWords Troubleshooting [SearchEngineWatch]

Blogged with Flock

Tags: GoogleadWords

Popularity: 79% [?]

There was a time when seemingly everyone was adding Digg-like voting to some of their products and services: Dell, Yahoo, Microsoft. Now, a bit late to the game, Google follows suit by adding user voting to their search results. The feature is currently in experimental phase and it may not ever be introduced to actual Google results.

The users can mark a particular search result as useful or remove it from their result list; the changes are permanent for that particular user, but they show only if you’re logged in.

Google Considers Adding User Voting To Search

Blogged with Flock

Tags: GoogleUser Voting

Popularity: 80% [?]