Why?

Sure they give you some free analytics to see how well your ad budget is doing. They also control how much each click costs you. Its not a far stretch to see that if I know you are making a $100 sale off of a $0.10 click, I need to charge you more for that click.

Can you say ‘Conflict of interest’?

I have received several mails recently inquiring how I have so many pages listed highly in Google, especially for tweaks, hacks and fixes for PHP and server related issues.

The key is the keywords used in the post title and body.

The keywords for any issue specific post, such as a bug or error message are easy. Portions of the actual error message should be used. Ask yourself, ‘If I had this problem, what would I search for to find the solution?’

Take for instance a recent post regarding an error I encountered after installing the Zend Optimizer for a client. The error message was:

Failed loading /usr/local/lib/ZendExtensionManager.so: /usr/local/lib/ZendExtensionManager.so: undefined symbol: zend_extensions

The post was titled: Undefined Symbol Error After Installing Zend Optimizer.

Checkout these Google results. Three top five links on a post that is less than two weeks old. Yes, those are some specific searches, but they were pulled from the server logs.

Some general guidelines:

1. Use distinctive keywords in page title and body
2. Use proper names where possible (application or company name)
3. Limited use of generic keywords in page title
4. Use whole phrases if applicable(error messages, features, slogans)
5. Reinforce keywords in page body
6. Add secondary keywords and phrases in page body

Many SEO boards went ballistic this week after a post by a Google employee let it be known that not only does Google allow it’s employees to have and use real Adwords accounts, but that they even encourage it. Of course their own employees don’t get rich with their inside information – ‘but not to worry, Google keeps an close eye on ‘em to make sure they don’t do anything unethical.’

The guy actually said this:

By policy, AdWords employees may certainly have AdWords accounts. Please rest assured, however, that they are thoroughly monitored and governed by a list of requirements as long as your arm – designed to ensure no conflict of interest.

It is perhaps worth noting that everyone concerned with AdWords, whether an engineer working behind the scenes, or the person that answers your email, is actually encouraged to have a small AdWords account. The purpose of this is to make sure that everyone who works on AdWords knows exactly what it feels like to be an AdWords advertiser.

[...]

Bottom line: Many Googlers have AdWords account. IMO, It’s a good thing, and it certainly is no secret. I can understand why this is concern to some of you, and I can also say with a great deal of confidence that you may rest easy

I checked in with a recent search engine marketing/optimization client yesterday to see how their sales were tracking. This client runs a bricks and mortar store and is building an ecommerce site to compliment the main store and create another revenue stream.

When I started working with them, they were doing about $40-$50 a day in sales. We implemented some optimizations in their site to increase their search engine ranking naturally, started an Adwords campaign and signed up a half dozen resellers to sell their products through their own sites and stores.

Three months in, daily sales had grown to $120/day. While we were working on their site we both spent time every couple of days monitoring sales, checking usage reports and Adwords reports and tweaking anything we could to maximize their potential at every step.

Yesterday (three months later) the client gushed how their natural search placement had been climbing last he checked, so he stopped the Adwords campaigns to rely soley on the search results for clicks. Afterall, it was free. Other than that, he said he hadn’t been monitoring things at all since shortly after I stopped working with them.

“What are your daily sales for August so far?” I asked.

$72/day

The reason for the significantly lower sales?

First, the amount of effort involved in improving the site had completed stopped. They are in a highly competitive (online at least) niche. Their two major competitors are masters of online marketing and are continually bettering themselves, leapfrogging over each other.

If you are in a competitive field with several players, you must spend at least a day a week to stay in the game with them, more time if you want them in your dust. Even a non-competitive field requires a couple of days a month to prevent losing all of the ground gained. Its not a 100m sprint, its a marathon. And its relative, you don’t need to be perfect, just better than your competitor.

Second, they stopped the Adwords campaigns all together. When I go to Google looking for information, I stick with the results in the natural search. When I’m buying, I look at the search results and the Adwords results. If a company has enough going on to be on the first page of Google and the first page of Adwords results, they must be serious about their business. This is perception, but when all you need is a click based on a two second decision, perception is everything.

Split testing allows a site to test two different approaches to acheive a goal. Set up both approaches (A and B) in a seperate page or email or post, split traffic equally between them and wait, then examine the results.

What’s your goal? Purchases? Email signups? Ad clicks?

Whatever the goal is, measure the responses for both A and B and when there is a clear winner, stop the test and send all of your traffic there. Declaring a clear winner can be a little tricky, but it will vary depending on your traffic and how polarizingly different your approaches are.

Now, take a break. Figure out a new approach to test and repeat. This way you will always be improving your response rate, and thus your bottom line.

Some variables to consider:

• Price points
• Length of content page (long and complete or short and to-the-point)
• Personal elements (signature, photos, testimonials) or direct business approach
• Headlines
• Call to Action
• Graphical layout / colors
• Location of each of the page/email elements

I have been shopping online a lot lately… really burning a whole in my check card. There are a lot of bad shopping experiences out there. There are sites with confusing checkout processes, sites with too many checkout steps, sites asking for too much information, sites distracting me from completing my purchase.

Here are a few of the easiest things you can do to streamline your checkout process.

Testing
Get your mom to buy something from you (unless your mom spends more money online than you do). The point is, have somebody you trust and who has little online experience, test out your site and actually make a purchase. Find out where they had questions, what was confusing and what they liked. You look at your site all day long, surely it makes sense to you.

Avoid Confusion and Distractions
Many high traffic sites try to get as much product in front of the consumer per pageview. That may work for getting more items into the cart, but once the user reaches the checkout process, it can be distracting. Is it worth risking of losing a sale by showing a “People who purchased this also purchased these” interstitial page inline while checking out. Sure, they may add another $5 or $20 item to their cart, but do you really want to give them more time to think about the entire purchase? He is already at the checkout counter with his credit card in the machine, don’t ask him to put it back into his wallet to go back and add a pack of gum. You may just lose him.

These distractions work well at the grocery store. But in ecommerce, clicking on the item takes the customer back out into the general store catalog. Now, she must make the purchase decision again to get to the checkout (twice in the same shopping experience!). Ecommerce is not like grocery shopping. I don’t have to pass by the checkout stand to leave the store, I can close the browser, take a phone call or go to another site at any time during the process, without buying anything. Don’t distract me at this most crucial stage of my shopping process. Streamline your checkout as much as possible.

Examine Where You Lose Your Customers
Examine server usage reports to see where customers are dropping out of your shopping experience. You can do this with almost any log tracking package if you follow these steps.

Write down all of the page names in your shopping cart software’s checkout process. They should be something like:

1. shopping_cart.php
2. address_details.php
3. payment_details.php
4. process_order.php
5. confirmation.php

Go through your checkout and write them down in order. The address and payment information capturing pages may be combined into one page if both of those actions occur on the same page of your shopping cart software.

Now go to your server log report and look up the number of pageviews for each of these pages over the last month.

Your list will look something like this now:

A graph of the breakdown is typically a funnel, pointing down from your shopping cart page to your final order confirmation page. The shopping cart page should have several times more pageviews than any others. This page is used during the normal shopping experience even without a user going to checkout. Some users will view their cart many times during the shopping trip just to see how much the damage is and if they can find out any other charges (ie. shipping and taxes). It may even be the resulting page in every add-to-cart action, to confirm what was just added to the cart.

The first page of the actual checkout (in this case address_details.php) will be the baseline for how many shoppers begin the checkout process. Now examine the chart and see if any of the pageview totals are way outside of the funnel. In my experience, each successive page in the checkout process should see somewhere between 25% and 50% less traffic than the page above it. If you see more customers abandoning your checkout process at each step than this range, you need to rethink your checkout. Do the math on the numbers and see what each step’s loss ratio is (ie. divide the number of pageviews of each page by the page above it). If one particular step exhibits a far greater loss ratio than any of the others, then that page needs optimized.

The object is to get the same number of customers that start the checkout process to actually finish it. One hundred percent retention through checkout is the goal, but it is unattainable in the real world.

Contact me to request an ecommerce consultation for your site.

Search engines tend to give high relavance to keywords in the domain name. Searching for ‘Pimp Costumes’? A domain with both keywords in the domain name is likely a good bet.

Best Tool For the Job posts a list of utilities.

DomainsBot is especially helpful because it accepts a list of words to spin into your next domain name.

How do you build inertia in marketing? The more people you can get behind an idea (positively if possible) the more the idea builds upon itself until it reaches a critical mass. Take BitTorrent from Bram Cohen. Started as a pet project he now lives off the donations from users. Not selling the software, DONATIONS.

The perceived value BitTorrent’s users have in the software is worth more than setting a price that a traditional software company would value their software at. Many won’t donate at all, many will donate what they feel is the minimum and still others will donate more than what a company would charge because of the ‘warm fuzzies’ (to use my high school coach’s favorite saying) they get from the software and the good will trust the author has placed in his user’s hands.

What’s one of the easiest ways to kill inertia? Fine Print. If you are trying to do something good for the consumer (ie. provide them with a quailty product and top-notch buying experience) why inundate them with pages of fine print? It kind of kills the ‘warm fuzzies’ doesn’t it?

Seth Godin posts in Tiny cuts some examples of how fine print and poor customer service can kill any marketing effort.

Fine print is everywhere I look. Fine print means that a lawyer has made sure that you probably won’t win a lawsuit, but is the lawsuit really the point?

When did marketers fall in love with the idea of overselling and then hiding, instead of doing precisely the opposite?

NewsFactor Network reports in Search Engine Choice Affects Results that which search engine we choose effects the view we see of the web. I had always known that results can vary greatly from engine to engine. In tracking my own sites its easy to see the differences between first page, second page and latter rankings. The most interesting thing in the article though, is the number of unique links when comparing two or more engines on the same search terms.

When the results of two search engines were compared, 11 percent of the hits appeared in both results tables. That overlap dropped to 2.6 percent when three search engines were queried and to 1.1 percent among four search engines. On average, 74 percent of the first-page search results were unique to Ask Jeeves. Slightly less, or 71.2 percent, of Yahoo’s first-page results were unique.

MSN search results were close to Yahoo’s with 70.8 percent of the first-page results being listed only on that search engine. Nearly 67 percent of the first-page results for Google showed no overlap compared to the other search engines.

Jason Calacanis offers a prize(?) to whoever can come up with a ‘kick butt’ list of the top 500 weblogs. $50,000 in advertising to anyone who could come up with a better ranking system or $10,000 cash to a programmer who creates such a system for his company.

To actually build a ranking system, you need the data. We need to build a bot to search and store all the metrics that we will rank on. While we’re at it, we might as well build a pinger that slurps posts as they happen.

Lets decide on the metrics…

• the text itself
• the domain/blog the post comes from
• trackbacks/pingpacks
• comments
• links from other blogs
• non-blogged links
• all the same info on each trackback/pingback

This also assumes that the best way to rank the most popular blogs is how many inbound links they generate. What about just controversial blogs and hot topics? They generate a lot of buzz because of the polarizing nature of the controversy. Does controversy or buzz actually equate to popularity? I could say something outlandishly pro-Bush or anti-Bush that gets picked up by an well read blog or a search engine and generates 100,000 pageviews in a day but does that mean I’m popular? If we add a dampening factor of time, inbound links sustained week after week over a couple of months we would find a more accurate representation of relative popularity. But then what about bloggers that only post a couple of times a week?

What other options do we have for rankings? The number of feed subscribers? The number of daily visitors? The number of top ten search results everyday? Anything we come up with can eventually be gamed by anyone who wants to. Why do you think Google’s ranking is such a closely held secret?

What, now, is the point of a top 500 blogs of all genres? Blogging is a largely personal experience, its off the cuff, emotional and more of a conversation. You can find blogs on every subject known to man. Who gains from a popularity list other than the ego of those involved? If I want to read from an authority on Italian cooking would the top 500 get me anywhere? Maybe one or two blogs. Lumping every subject together in one big list thorws away one of the most important advantages of blogging compared to other forms of media… the individuality of subject.

Now we need a topic-based parameter.

Take the data gathered from this pinger/bot and build an interface that lets me manipulate the results like a fast food menu. ‘Give me 250 of the most actively linked blogs on PHP coding over the last 45 days’. Now thats a list that would be useful to me (today). Tomorrow could be a whole new list.

Give me some fancy buttons to post on my blog linking back to my top list and every other list out there would be playing catch up for the next year.

Screw the $10,000 I want in on the backend profits.