Posted: January 5th, 2010 | Author: admin | Filed under: seo, social media | Comments Off
A few days ago an Aardvark user asked me a really common question: how should I promote my new website?
The web is awash with people claiming to be “social media experts” and “SEO consultants”. Their advice in my experience ranges from the unhelpful through the inaccurate all the way to the positively counter-productive. Many of them are snake-oil merchants who want to charge thousands of dollars to dispense obvious advice or to use shady techniques that will get your site blacklisted by the search engines for spam.
So here is my own obvious advice, for free. My credentials are more than a decade’s experience building websites large and small and a working knowledge of modern search engines*, but much more importantly that, I am not trying to get you to pay me.
For the sake of this discussion I am assuming you are a low- or zero-budget site; if you have a lot of money just buy an ad during the superbowl.
In short, there are three really good ways to make your website popular:
- making it easy to find your site using a search engine (a process called Search Engine Optimization, or SEO)
- interacting directly with potential users on social networks and other websites that allow person-to-person communication (known as Social Media Optimization, or SMO)
- organic growth (also known as Being Awesome)
Let’s tackle these in turn.
Search Engine Optimization
There are two facets to SEO: I have previously covered making it easy for the software used by search engines to understand your website. The other facet, and by far the more important one, is improving your inbound links.
By far the best way to get traffic to your site is to get other sites on the same topic to link to you. It gets traffic directly — by visitors on those sites clicking the links — and it also gets you placed higher in search results. No amount of tweaking and fiddling will get around this fundamental fact: Google (and all other major search engines) decides how relevant your site is by how many other people link to it, and how relevant they are. So the best place to promote your website is on other websites about the same topic. But it’s important to do this in a friendly, respectful and above all non-spammy way. “Non-spammy” is tricky, but some good guidelines are: not too often, not too fast, not to too many people at the same time, and not to an audience that won’t appreciate it.
Getting links to your site
- Just ask. Find sites on similar topics and drop the owner an email asking if they’ll link to you. Offer to link back to them in return; this makes you both look better in the eyes search engines. If they refuse, that’s fine — move on, and don’t be pushy. Remember to contact each site individually — a mass email to fifty website owners is going to get flagged as spam, because it is.
- Forums. Find places where people are talking about the subject of your site, and join in. Ideally, find a conversation where you can contribute something genuinely valuable to the conversation, and mention the name of your site or a link to it from there. Don’t be coy about what you’re doing — people know spam when they see it. “I hope
you don’t mind a little plug, but I have a site about this topic,” is acceptable. Do NOT just join a site, post a single message about your website, and then leave — that’s just spam.
Some forums also have a “signature” that you can append to your messages, and it is sometimes considered acceptable to include a plug to your own site in there — but be careful, as other sites will consider it spam. As usual, the key is spending some time and getting to know the culture of the site.
- Paid links. Unlike the other two this costs money, but can also be very effective. Find a highly-rated site and pay them to link to you (it’s best practice that they make it clear that it’s a paid link). Lots of smaller sites are happy to do this and even happier for the revenue, and it can be very cost-effective.
These are simple techniques, and require a lot of time and effort. However, they are also the most powerful. A single link from a popular site can mean thousands of visits, which could cost you hundreds of dollars if you were paying for ads.
Social Media Optimization
The very phrase “social media optimization” reeks of sleaziness. Social media is much harder to get right, but also potentially more powerful: a single well-timed mention of a topical page on your website can go around the world, spreading virally to millions of people in just a few hours.
The two biggest social media channels are Facebook and Twitter, but these techniques can be adapted for almost any website where people have profiles and send messages to each other.
On Facebook
Create a Fan Page for your site. Use it to describe your site, and to mention useful, relevant content from your site. Encourage your existing users to add themselves as fans: that event turns up in their news feeds, which their friends see, and is a great way to spread the word about your site virally. People tend to have friends with similar interests, so the rate of response to a mention in a news feed is much better than an ad, no matter how well targeted.
You can use your fan page to send broadcast messages to your fans (though these are often ignored if they are too frequent, so use them sparingly). You can also post messages to your Wall, which is a great way to keep your page fresh and interesting to new fans.
On Twitter
Create a profile for your site. As with your Facebook Wall, tweet useful information about your topic. Uniquely to Twitter, you can use the ad-hoc conversational dynamic to your advantage: search for your topic, and you will find people talking about it. If you have something useful to add — and only then — you can reply to them publicly (an “@reply”). If they like your tweet, they’ll follow through to your profile and find your site. If it’s particularly helpful, informative or witty, they may even re-tweet your message, sending it out to all their followers.
@replies are really powerful, but again, don’t be spammy: don’t reply to everyone with the same generic tweet, and don’t blindly reply to anybody who mentions your keyword — make sure they’re actually interested in your topic.
Avoid viral gimmicks like “Retweet this to win prizes!” These get you a whole lot of publicity, but at much lower quality: these people aren’t genuinely interested in your topic, they just want your free iPod. You also run the risk of annoying potential users who receive the same tweet from ten different people.
And finally, the most important and powerful technique of all.
Organic Growth, aka Being Awesome
One of the oldest maxims of website promotion is that content is king. The Internet, in all its marvellous chaotic complexity, is really good at surfacing quality content. So make sure your site is well-written, easy to read, and frequently updated. The best way to get popular is to be genuinely useful, informative, or entertaining, and no amount of futzing around with social media is going to get around that
It also helps if your site has a reasonably narrow, defined topic or range of topics — if it’s “my blog about everything” it’s going to be difficult to get anybody but your closest friends to read it, no matter what you do. Work out who you want your audience to be, and then come up with the most useful content you can possibly think of for them. They will reward you with real, long-last reputation and traffic.
Posted: June 4th, 2007 | Author: admin | Filed under: accessibility, search accessibility, seo, webdev | Comments Off
In my last post, I covered the politics of search accessibility, and why making your site available to all users is above all the profitable thing to do, without considering whether it’s the right thing. So now I’m going to cover how to make your site search accessible.
Please Feed the Spider
The program that runs around the Internet reading every single page and throwing it into Google’s* giant database is GoogleBot (Yahoo!’s is called Slurp). GoogleBot is your best friend, your worst enemy, your teddy bear and your mommy all rolled into one. GoogleBot is a very, very clever piece of software, but it’s not magical. Here is what GoogleBot does:
- It reads the text on your page and looks for “important” words and phrases
- It reads the links on your page and sees what pages you’re linking to
- It reads the links on the rest of the Internet and looks for pages that link to you
- It then calculates how relevant your page is to the words in it, based on the words on the pages that link to it, and how relevant they are based on other sites, and so on
Key take-home: it’s all about keywords and links. It is all about text. Attractive design and a witty site slogan and pictures of bikini models holding your product count for naught. As I mentioned in my last post, Google is in effect a disabled user using only the most basic of assistive technologies:
- It cannot see your images
- It will not execute JavaScript. Not any. (Real disabled users can often do JavaScript using better software these days)
- It’s not reading every bit of text on your page. It’s looking for the important words. And it’s in an almighty hurry.
- It does not follow links that do not look like web pages.
- It does not magically work out what your site is about. You need to make it obvious.
Already, some of the key things you need to do for SEO are obvious, in order of importance:
- Link text is important. Every time I see a link saying “click here” in 2007 it makes me want to weep. Link text is, above anything else, how Google decides what the page you’re linking to is about, and by working out what you’re linking to is how it works out what your page is about. Outgoing links aren’t silly, they’re essential.
- All your information must be accessible with JavaScript disabled, because that’s how Google sees the page.
- All your links must go to real information. “#” links are ignored, as are “javascript:..”. And the information on those pages must be relevant to the link text, obviously: don’t just link back to your home page.
- Any images must be described as text somewhere on the page, either within the ALT attribute, or some other technique**. Google ignores images (even image search is based on the text nearby).
What’s an Important Word?
It’s important to know what Google considers an “important” word. Google is more than a little secretive about this, but Google has its own guidelines for site design and professional, non-evil SEO people have their own search accessibility guidelines. My own, very subjective impression from several years of experience, is that the most important words on your page are:
- The link text going to your page. Nothing you can do about this but be a very good website, and hope people link to you. You can do your bit by linking to other people with sensible keywords, of course, and hoping they link back — but trading links explicitly is something GoogleBot is designed to detect. And it’s been spotting fakers a lot longer than you’ve been faking, so I don’t recommend trying to fake it.
- The page title. Don’t repeat your site name and slogan endlessly: say what this page in particular is about. Put keywords in there! It’s also what users see on the search results page, too, so make sure it makes sense to human beings.
- The meta description tag. This is an odd one: Google doesn’t pay too much attention to it in calculating relevance, but at a certain level of relevance, this is what it puts as the text under your page title in the search results, where it suddenly becomes very important to users who are about to click your link. So it’s important that this text is descriptive, useful, and short — something under 100 words. And again, load up on keywords. Repeat yourself, phrasing the same thing several different ways.
- H1 and H2 tags. H3 is dicey and everything beyond that is meaningless, but H1s in particular are super important, but only because they are rare. If everything on your page is a goddamn H1 obviously Google is going to ignore you. Use 1 or 2 H1s, and less than 10 H2s.
- ALT attributes on images. This is way down the list, so if you have really important text in your images, it’s best to use the technique I outlined in the third footnote so that it turns up in an H1 or H2.
Order is important, or, Don’t use Tables
Another aspect of your page that is extremely important to Google is source code order: literally, the order things appear in your source. Things that appear early on are likely to be more important than things that appear later. That seems obvious, right? But now look at your code: you’ve got the head, full of juicy meta data, and then you’ve got 5k of navigational elements, sidebar text, various other cruft, just placed first because you were using a left-floated column and so it was easier to put it there. This is killing you.
What’s much worse is when your source code order physically separates content that is semantically related: for instance, your headline is at the top of your page, then you have 5k of navigational cruft, then you have your content. Google will either fail to realise that your headline is describing your content, and thus not link the words, or worse, it will decide that your page doesn’t actually have any content on it relating to your headline, and you’re trying to spam it. Danger, Will Robinson!
And of course the number one offender from this perspective is using tables for layouts. If you care about web development, you’re probably aware that tables have serious issues with flexible, attractive layouts. However, that’s usually not a good enough reason to take to your boss: after all, it doesn’t bother her that your job is hard. However, tell her that using tables is causing an 80% drop in traffic to your site (as I explained in the last post) and suddenly you have an easy, obvious business case for reworking the layout of your code.
Tables put data into grid layout. If your data is in columns — and it frequently is, this means you often end up with a site code layout that looks like this:
| Site logo |
Article headline |
|
|
Article body |
To Google this reads like:
- Site logo
- Article headline
- List
- Of
- Nav
- Links
- Article body
So you can see why Google might get confused. So examine your code, and put things in the order of importance: you can use CSS to move stuff around on the page later. Coincidentally, source code order is also the order in which screen readers will read out your page to a blind user. So once again there’s a useful coincidence of making your site accessibile when you make it search accessible.
Of course — and I would have thought this was obvious, but I get questions about it that indicate to the contrary — you can use tables when the data is tabular. Don’t try to mark up your spreadsheet data using a series of stacked lists. Tables have real semantic meaning, but it has been diluted almost beyond help by their consistent misuse.
There is more I could tell you about SEO — the various hazily-defined statistical rules about how many links on a page is too many, optimal keyword density, and more, but these advanced techniques are icing on the cake, and the cake is made of search accessibility. It doesn’t matter what your keyword density is if Google can’t even get to your pages. So get out there and make the case for accessibility. And when the traffic is rolling in and your boss is giving you your huge bonus, you can get a tiny little extra bit of joy from knowing your site is also accessible to disabled users.
Posted: May 31st, 2007 | Author: admin | Filed under: accessibility, search accessibility, seo, webdev | 2 Comments »
So I was at the @media America conference last week. There was much talk of accessibility and how to do it properly, when to do it, and even when not to do it. There was also talk about why to do it, but that’s where I think the speakers dropped the ball. Accessibility is not about helping disabled people: it’s about money, and you making more of it. (I’m going to use a lot of bold text in this post to emphasize stuff. That’s because it’s long, and you’re skim-reading. See, I know you.)
Accessibility: because it’s the Right Thing™?
The dirty secret of accessibility, swept under the rug by many an evangelist, is that the cost of making your site accessible is relatively high: in my experience, something like 20% additional dev time on a new project, although experienced developers can bring this down, and the cost decreases dramatically for incremental updates once the project is up and running. But a 20% margin is definitely non-trivial. And if you’ve not been thinking in terms of accessibility from the start, this pricetag rises sharply: retrofitting accessibility often involves fundamentally reworking the architecture of a web page*. You’ll be looking at spending something like 50% of the time you spent originally developing the site on the retrofit. Ouch.
The other dirty secret of accessibility is that the number of disabled users is relatively low. Not tiny, but I often hear figures like “60% of Americans are disabled”, and while this is true, it’s disingenuous because that figure can include people like amputees or paraplegics who can use the web with no problems whatsoever. The truth is that somewhere between 10% and a maximum of 20% of your users will have trouble using your site without assistive technologies. This makes it a very close call, when starting a new project — serving 80% of your possible users doesn’t seem ideal, but is an acceptable loss to get it out of the door 20% faster, right? You can build the accessibility in later!
Except you can’t. After launch, you’ve got an inaccessible site and you’re facing a 50% dev time bill to retrofit that acessibility in: another 3 weeks on what was a 6-week project, just to get 20% more users? That makes no business sense: much better to build another project, and get another 80% of users in the door quickly.
This is the kind of unavoidable math that has made the web inaccessible today. And that’s the harsh truth: building in accessibility for disabled users does not make business sense. It’s still a good idea, a noble idea, but it’s not a financially sound one. This is true in the real world, too, which is why legislation was necessary to force everybody to put accessible toilets and wheelchairs in everywhere.
Accessibility: because you could get sued?
Of course, legislators have (eventually) worked out this problem, and as such there is already web accessibility legislation in place in many countries that makes it illegal to produce an inaccessible website. Problem solved! It’s the law! We have to do it! Right?
In an ideal world, yes. In the real world, the law is only patchily enforced. Only a few very large, very high-profile sites have been sued so far (plus some government sites). You can always fly under the radar, hope nobody notices, and not build in accessibility until they sue you. It’s a good gamble to make to avoid increasing the cost of your site by 50%, right? Again, the math defeats us.
But this is all very unsatisfying. You, the clever, compassionate, standards-compliant modern web developer, feel that this cold logic is intrinsically, morally wrong. So you make the case for accessibility: you try to inflate disabled user numbers (counterproductive; it will make your manager trust you less) and deflate the amount of time it will take to make it accessible (an even worse idea; now you’re missing deadlines because of “that damn accessibility stuff”, making your manager hate the whole idea).
So here’s how you, as a developer, can stay true to your noble impulses to build an elegant, accessible website: stop calling it accessibility.
SEO: Open up, Google is coming!
Search Engine Optimization, or SEO, is the hot shit right now. Google is the Internet for a lot of people, and if Google can’t find it, then it doesn’t exist. Nobody goes deep-diving on a site to try and dig up information anymore. Either they type in their search terms and your site comes up with exactly what they need on that page, or they will never click your link. Sites these days get 50-90% of their traffic from search engines**, and the overwhelming majority of that is deep links to pages within the site.
So it’s absolutely imperative that search engines be able to access your site, and this isn’t just keywords on your home page: Google must be able to get at every single page of the site, every nook and cranny, and see every little bit of information. A site that can’t be indexed is throwing away up to 90% of its audience. In other words, this traffic is lost by sites that are not search-accessible. And there’s an interesting word in that phrase.
Search Accessibility: because you’d be an idiot not to
Here’s the final dirty little secret of this situation: Google is a disabled user. Or more accurately, Google has all the same limitations of somebody using assistive technologies:
- It doesn’t look at pictures
- It does not execute any Javascript
- It isn’t reading every bit of text on your page; it is looking for the important bits
Suddenly, the equation changes: at least 55% of your users need your site to be accessible, and possibly over 90% do. Only 10-20% of them need it to be accessible all the time, but that doesn’t matter, because up to 90% of your users will never even visit your site if it isn’t search accessible. This isn’t out of solidarity, or legislation. They simply won’t find it. Search accessibility is not an optional component, to be bolted on after the main launch. Chances are, if you haven’t got your search accessibility right, there will never be a second launch, because your site will fail.
How can I further underline the importance of search accessibility to a web-based business? Let’s turn the numbers around: you can more than double traffic to your website by making it search accessible. Does that sound like something you could take to your manager as a business case? Keep in mind, 50% traffic from search engines is an absolute minimum. If you’re getting 90% of your traffic from Google, then making yourself search accessible will result in a tenfold increase in traffic. Those sorts of numbers are why SEO is now big business, with a whole industry built around paying consultants to tell you how to get it right. That industry wouldn’t exist if they weren’t getting results.
But you don’t need to pay somebody. Once you’ve got the big, obvious business case out of the way, and swallowed the bitter pill that doing things properly will take 20% longer, search accessibility is super-easy. For my own personal how-to for search accessibility, see my next post.