Put another way, this means that webmasters can now provide Google, Bing, Yahoo and Yandex with much more granular information about products and offers on e-commerce sites in a manner that is officially sanctioned by these search engines. In terms of e-commerce SEO, this is potentially a pretty big deal:  it is a means of providing very exact e-commerce information to the search engines in exactly the form they want it.

The Integration Of GoodRelations Into Schema.org

As Barbara Starr emphasized in a previous article, GoodRelations and schema.org are both vocabularies, and should not be confused with the syntaxes available to encode these vocabularies. Both of these vocabularies can be marked up in HTML using either the RDFa or microdata syntax (although initially, schema.org was very much geared toward microdata).

Until now, these were related but separate vocabularies. With this integration, however, the bulk of the GoodRelations vocabulary has now been made available in schema.org.

For those who had been using RDFa to markup GoodRelations, the situation is unchanged. The developer of GoodRelations, Martin Hepp, said in a post on the technical background of the integration with schema.org that “GoodRelations will remain an independent vocabulary, and usable in RDFa and other RDF syntaxes.”

The majority of sites using schema.org, however, have been employing microdata, including e-commerce giants like WalMart, Overstock and eBay. For these sites, the type of e-commerce information that can be marked with microdata has been vastly extended.

New Product Types & Offer Properties In Schema.org

While the integration of GoodRelations has resulted in a multitude of new types and properties being made available in schema.org, the types most obviously extended by GoodRelations are schema.org/Product and schema.org/Offer.

The schema.org Product (left) and Offer (right) classes before and after integration with GoodRelations

The number of Product-specific properties have grown from eight to 25, and three product-specific types have been added where previously there were none.

The additions now make it possible to describe physical properties of a product, allows relationships between products to be defined, and extends the number of product identifiers (like SKUs and part numbers) that can be associated with an item.

• Product physical properties color, depth, height, weight, width, itemCondition These properties now make it possible to provide structured data about the physical dimensions, appearance and condition of products.
• Product relationships isAccessoryOrSparePartFor, isConsumableFor, isRelatedTo, isSimilarTo These properties now make it possible to formally link related products. Previously, there was no way in schema.org to indicate which camera a lens cap fit onto, or what battery that camera required: now, these relationships can be explicitly declared.
• Product numbers gtin13, gtin14, gtin8, mpn, sku These properties allow much more specific product identifiers to be declared than the previously all-purpose property productID.
• Specific product types IndividualProduct, ProductModel, SomeProducts Allows everything from very specific products (such as a laptop with a specific serial number) to very general products (such as multiple, similar laptops) to be described with Product properties.

The number of Offer-specific properties have grown from ten to 37. These additions make it possible to provide much more detailed information about prices and offer conditions, as well as allowing vendors to describe e-commerce offers other than the sale of physical products.

• Offer conditions acceptedPaymentMethod, addon, eligibleQuantity, eligibleRegion and other properties These properties facilitate the description of all types of offer conditions, such as the payment types that the vendor accepts, or the geographical regions for which that offer is valid.
• Offer availability availabilityEnds, availabilityStarts, availableAtOrFrom, availableAtOrFrom, availableDeliveryMethod These properties make it possible to state the length of time an offer is available, where the offer (usually, but not always, a product) can be obtained, and what delivery method is available for the offer (which can be everything from a specific delivery service to a computer download).
• Offer business function The property businessFunction and its expected type, BusinessFunction This allows markup of offers besides simple offers to sell, representing a huge extension of types of transactions supported by schema.org. Apartment rentals, car leases, maintenance agreements and even offers to buy can now be marked up.
• Offer warranties The property warranty and its expected type, WarrantyPromise The duration of a warranty and the precise scope of services a warranty covers can now be explicitly stated with schema.org.

These are cursory descriptions of just the main additions to Product and Offer in schema.org as a result of GoodRelations integration.  E-commerce site owners would to well to examine the Product and Offer classes in detail to determine which specific properties might apply to their products and services (keeping in mind that what constitutes an e-commerce site, from a schema.org perspective, has been vastly extended).

As well, the impact of GoodRelations integration extends well past these two e-commerce-specific classes. OwnershipInfo, for example, provides a way of formally describing the ownership history of a product.

OpeningHoursSpecification supports a more granular way of specifying hours of operations, as well as making hours of operation applicable to more things, such as places. And so on… Again, there’s a lot to discover for anyone not previously familiar with the GoodRelations vocabulary.

Rich Snippet Benefits

Functionally, the chief benefit of the integration is that it extends the number of things that you’re able to describe to the search engines using structured data markup, especially things in the realm of e-commerce.

Currently, the chief SEO benefit of providing such markup is increased visibility in the SERPs when search engines generate rich snippets for a product in the search results. Here’s a rich snippet in Google for a camera on eBay, which uses schema.org Product, Offer and Review markup.

It’s still far too early to see if the search engines will produce rich snippets for any of the GoodRelations types and properties newly integrated into schema.org (things like “For rent from $623 per week” or “1 year warranty”). However, I’d be very surprised if we don’t see more product and offer information appearing in the SERPs based on the extension of the schema.org vocabulary.

Certainly, e-commerce sites desirous of a competitive advantage won’t want to wait until new rich snippets start appearing before adding new markup to their pages. And businesses whose offers weren’t adequately addressed by the original schema.org vocabulary will certainly want to add markup if that now allows them to provide detailed structured data to the search engines about their products and services.

Rich Snippet & Structured Data Testing

As reported by schema.org’s Dan Brickley, Google’s Structured Data Testing Tool “was updated before we pushed the new site.”

Accordingly, the Testing Tool should now validate or provide error messages for code using the extended schema.org vocabulary.

The Testing Tool certainly didn’t have any problems validating an example snippet contained in Martin Hepp’s post detailing the integration, although testing hiccups are likely to occur during the initial phases of the roll-out.

SEO Benefits Beyond Rich Snippets… & Benefits Beyond SEO

Will supplying even more specific e-commerce information to the search engines result boost a site’s organic rankings? However, it stands to reason that sites capable of correctly offering more extensive e-commerce information will by necessity be well-structured, rich in detail and technically sound — attributes generally regarded as favoring higher rankings.

Outside of the realm of SEO, this integration has potential repercussions for the entire ecosystem of commercial transactions on the Web. Search engines are, after all, not the only data consumers that can digest and make use of semantic markup.

OpenLink founder Kingsley Idehen openly pondered the implications of the integration and said, first of all, that “structured data driven e-commerce at Web-scale is now a reality — think Amazon.com on steroids as a natural feature of the Web.” It also facilitates serendipitous discovery of offers, wish-lists, products etc., and does all of this without compromising privacy.

In short, it brings us closer to the realization of the linked open commerce model that he, Martin Hepp and Aldo Bucchi have been working on since 2009. The model envisions the interlinking of e-commerce platforms, big Web retailers, product catalogs and ontologies — with GoodRelations providing the universal e-commerce model.

Pretty geeky, right? Fortunately, the site also provides a very concrete example of where linked open commerce might take us.

Even if this integration doesn’t usher in a brave new world of intelligently linked e-commerce information, it certainly does allow e-commerce webmasters using schema.org the opportunity to provide more detailed information to the search engines about what’s offered on their websites.

Linked open commerce images above used under a Creative Commons license from linkedopencommerce.com.

The mostly widely cited and easily verifiable of these benefits is the generation of rich snippets:  a specially formatted search result block that includes information specific to the type of resource being referenced.

Might there be other SEO benefits to using structured data?

On one hand, Google has insisted that the advantage of employing structured data (in the form of structured markup) is limited to the production of rich snippets. On the other hand, bald claims have been made that structured data use, in and of itself, improves the ranking of Web pages in search.

I believe the truth lies somewhere in between. While employing structured data by no means guarantees superior rankings, the provision of metadata can potentially provide the search engines with a better understanding of what any given Web resource is about.

As I’ll argue, the search engines would have to willfully ignore in their ranking algorithms information they ascertained was reliable enough to return in a rich snippet. This is possible but (I think) highly unlikely.

Attribute-Based Structured Data

The most widely employed use of structured data for search visibility are mechanisms that rely on HTML attributes to encode metadata in web pages, usually referred to as structured markup.

With structured markup the presentation layer (what humans see when the look at a Web page) is separated from the data (what computers see).

In this way, search engines do less guesswork because they receive explicit information about elements on a page.

For example, 10:12 is less likely to be misinterpreted as a ratio when it is marked up in a time field.

The chief types of structured markup now widely supported are:

• Microformats - Structured markup for very specific topical realms, such as recipes (hRecipe) or people (hCard).
• Microdata - Structured markup based on HTML5 attributes and – in terms of official search engine support – referencing the schema.org vocabulary.
• RDFa - Structured markup that can draw on any number of vocabularies, making it very extensible but also more complex than microformats or microdata.

Any of these attribute-based markup formats are capable of producing rich snippets for currently supported types, such as online product offerings, reviews and events.

A recipe rich snippet from Google and a product rich snippet from Bing

But do they also facilitate better search engine visibility, either in the form of higher rankings in the SERPs, or the inclusion of pages in result sets for which they might otherwise be excluded?

Google asserts that “marking up your data for rich snippets won’t affect your page’s ranking in search results.”

But on the same page, we’re also informed that:

“Providing this information doesn’t affect the appearance of your content on your own pages, but it does help Google better understand and present information from your page.”

That structured markup helps Google better “present information” is, of course, self-evident insofar as they encourage its use for rich snippet generation.

What is more interesting is the statement that it helps Google to “better understand” content.

Is it not reasonable to think that when your content is better understood this might lead Google to include or elevate that content, at least where that improved understanding helped Google to better associate that content with a relevant query?

Bing ‘s article on annotating structured content provides a similar insight (without, as far as I can determine, speaking explicitly to the question of rankings, one way or the other):

“Annotating your data doesn’t actually change the visible content, but gives Bing valuable information on the type of content you’re hosting on your site. On our side, we put your annotations to good use, for example by using them to increase the visual appeal of your search results, or to supplement and validate our data sources.”

Like Google, Bing suggests here that the value of structured markup may extend past rich snippets (“the visual appeal” of search results), but also serve to provide additional information about the annotated content in general.

Again, it seems to me that better understood content, or content that might be deemed more trustworthy by dint of seemingly copasetic structured annotations, could at least potentially fare better in the search results for relevant queries.

I’ll return to this thought after looking at another type of data annotation by way of comparison:  metadata rich XML files.

“Structured” Data:  Beyond Attribute-Based Structured Markup

The search engines (and Google in particular) have long supported a number of different protocols that provide additional information about a resource directly through the provision of an XML file. Sitemaps and RSS feeds are what I’ll use as examples here.

To a semantic Web purist, these data types are not “structured data” because they are not, strictly speaking, related to the Resource Description Framework (RDF).

However, insofar as the search engines understand and readily consume these XML formats, they are capable of deriving the same sort of information from them as “true” structured data. In short, they provide the search engines with metadata about URIs.

Just as the benefits of structured markup may extend beyond the appearance of rich snippets in the SERPs, the benefits of sitemaps and RSS may extend past their traditionally-extolled virtues related to indexing and subscriptions, respectively.

XML sitemaps, at a basic level, allow webmasters to tell the search engines about the URLs available on a domain, and provide the search engines with hints that may expedite the re-indexing of important or frequently-changing pages.

However, they also allow webmasters to provide the search engines with more detailed information about a specific URL, including information that might not be available on the webpage itself.

Consider these examples of tags for different types of sitemaps and Google’s description of each.

• Video Sitemap  <video:tag> – A short description “of key concepts associated with a video or piece of content”
• Google News Sitemap <stock_tickers> – The “stock tickers of the companies, mutual funds, or other financial entities that are the main subject of the article”
• Image Sitemap <image:geo_location> – “The geographic location of the image”

These tags all extend the value of sitemaps beyond ensuring or expediting indexation. They provide the search engines with additional information that may enable the resource to be displayed in the results for relevant queries.

The net impact of this, from an SEO perspective, is that videos, pages and images may appear in the SERPs on the basis of the metadata that has been provided.

In this way, an image tagged in an image sitemap with “Dublin, Ireland” may have a better chance of appearing in results for “photos Dublin” than the same image on another site lacking that geographical information.

Similarly, assigning a stock ticker symbol should certainly help Google ascertain whether or not a news article pertains to a corporate entity. If there’s any doubt in Google’s mind as to whether occurrences of “Apple” in a post refers to the company or the thing you eat, adding the markup “NASDAQ:AAPL ” should remove any ambiguity.

This is not to claim that employing sitemap tags is in itself sufficient to gain visibility in the SERPs for relevant queries. Google will certainly rely heavily on the context of the page in which an image appears to determine its geographical relevance, and it can almost certainly tell the difference between a fruit and a computer manufacturer in the absence of metadata.

But, if these types of sitemap-provided data do not at least potentially provide Google with a better understanding what a resource is about, then it strikes me as curious that they actively solicit this information.

The brief point I’ll make about RSS feeds are that they are conceptually similar to XML sitemaps in that they provide metadata that is consumed – and used – by the search engines (a fact often overlooked by webmasters that singularly regard RSS only as a subscription mechanism).

I once hounded a blog I read regularly to differentiate their site-wide <title> tag, although I was unable to cite their presence in the SERPs as a reason, as Google was returning the RSS title <item> in search snippets rather than the <title> tag itself. I won’t speculate on whether or not the provision of RSS aided in the ranking of their posts, but it unquestionably had a favourable influence on the visibility of those posts in the SERPs.

Visibility In Search Beyond Rich Snippets

When you use structured data, you provide the search engines with a better understanding of a resource. If employed successfully, structured data shouldn’t invoke a “uh-huh” but an A-HA! moment from the search engines: not “yeah, that’s a product price – so what?” but “that’s a product price – good to know!”

Just in the same way that traditional HTML optimization is beneficial because it provides the search engines with more information about a resource (an <img> ALT attribute, an <h1>-qualified title, and so on) so does structured data.

In the same way that a Google News sitemap can potentially help Google tell the difference between Apple and an apple, structured markup might also help in situations where disambiguation is an issue.

If you’ve told Google and Bing that your page is about the book Tom Jones rather than Tom Jones the performer, wouldn’t it stand to reason that the former would have a better chance of appearing in search results for the query “tom jones fielding” and the latter in results for “tom jones musician”?

So using structured data may not “get you high rankings” in the sense that links or lots of quality content might, but as with other metadata it absolutely has the potential to be beneficial.

If you’re a typical SEO you’ll optimize the <title> tag to make sure it contains important keywords and is descriptive of the page’s content. Does this guarantee that your page will have better visibility for a target search term? No.

But you have provided the search engines with further information about your page, that they’ll check against other data points and probably use if its valid – which is why a web page with a descriptive <title> tag typically outperforms a similar page that lacks one altogether.

The principle applies to all sorts of optimization techniques and strategies, such as optimized anchor text, flat information architecture and breadcrumbs.

These are all things that, like structured data, allow Google to better determine the topicality of your pages and your site, but themselves don’t guarantee “high rankings” simply because you’ve employed them.

But a journey down the conversion funnel always starts elsewhere:  the display of a banner ad, the entry of a search query, the appearance of a link – even a blank browser window poised to accept a type-in address.

In short, not all traffic to a website is created equal, and these individual sources must be taken into account when crafting strategies to increase conversions.

Targeting Specific Sources Of Traffic

Which traffic sources are likely to generate the highest conversion rates? According to one case study, paid search is the most effective converter. Another cites social media referrals as a conversion hero. Still another clearly shows that direct traffic outperforms any other source of referrals.

There are numerous reasons why conversion benchmarks by traffic source are unreliable. Obviously the nature of the target website plays a role in conversion. Established ecommerce brands like QVC or Amazon will almost always perform better than relatively unknown sites, even if they sell similar products at a similar price point.

The specifics surrounding a particular source of traffic are also important.

An email campaign targeting previous customers who have specifically opted into a mailing list will have a much higher success rate than an identical campaign blasted to 100,000 random email addresses purchased from a list vendor. A high-budget paid search program targeting only branded keywords will likely do better than one making low bids on generic keywords.

The devil is in the details, so don’t dismiss out of hand any given traffic source because of a supposedly low conversion rate, just as you shouldn’t aggressively pursue purportedly high-converting traffic that may turn out to expensive and ineffectual. Through trial and error you’ll be able to uncover traffic sources that convert at a reasonable rate for a reasonable price.

This is not as simple as spending more on those sources which offer the highest conversion rate, however. In calculating return on investment you must also account for the cost of the traffic and the average revenue that each conversion from that source brings in. So display advertising, for example – a notoriously bad converter – may turn out to a profitable traffic source for your website because the cost of impressions is so low, and the volume so high.

Matching The Message To The Medium

Insofar as it’s feasible, you should endeavor to direct traffic from different sources to an appropriate location on your website. Whether that location is a specifically designed landing page or an existing page on your site, the message on your site should mirror the messaging of the off-site traffic source.

It’s not always possible to send different types of traffic to custom landing pages. But it is often possible to change the message that appears on a single page based on the source of traffic.

Perhaps the most widely-known example of this strategy is dynamic keyword insertion (DKI) for pay-per-click campaigns (I’m speaking here of dynamically inserting search keywords on the landing page, not to be confused with dynamically inserting keywords into the PPC ad itself). The offer on the landing page remains constant regardless of the keywords used to discover the page, but the message on the landing page is changed to include references to those keywords.

Below I describe a number of conversion strategies for specific types of traffic, but often, variations on this sort of mechanism can be used to increase conversion rates. Whatever the source of traffic to your site, you’ll have a much better chance of converting visitors if you use customization to better meet their expectations.

Optimizing Traffic From Organic Search

While it might not seem initially that you can do much to boost conversions from organic search, it is certainly possible to align the messaging that appears in your search result snippet with what the visitor sees when they click through to your site.

The title and meta description of a page is usually what appears to visitors in search results, and the message they see once they click through to the target page should be closely aligned with the promise, if not the actual wording, of the search result snippet.

Your search engine optimization efforts themselves should anticipate the relationship between organic search traffic and conversion. Both in regard to external link development and internal linking, SEO efforts should be geared toward having the right page rank for the right term.

Build deep links so that the appropriate target page appears in the search results for any give term. If you’re successful in ranking for “blue widget” the search result should link directly to the blue widget page, not to the home page where the blue widget is one or more clicks away.

Optimizing Traffic From Paid Search

In many cases, employing a landing page that reinforces and logically progresses the message of a paid search ad offers the best chance of converting paid search traffic.

Aside from the exact control over messaging that a landing page provides, it also provides the opportunity to test different elements of the landing page in order to improve conversion rates through the life of a paid search campaign.

In the absence of a specifically designed landing page, the target page for any given pay-per-click ad should at least be closely aligned with what’s advertised. For an ad that matches the keyword query “Nikon camera” ensure the visitor lands the category page for Nikon cameras – not cameras in general, and certainly not the home page.

While this is very much conventional wisdom it bears repeating, as it is still extremely common to see ads centered on very specific keywords link to a single, generic click-through location.

Optimizing Email Traffic

Like paid search, custom landing pages are likeliest to perform best for email campaigns, though like paid search this is dependent on the details of the marketing message. Unlike paid search an email may present links to a number of offers, and it’s perfectly acceptable to link to, say, individual product detail pages if that’s what the email is advertising.

Whether the click-through targets of an email campaign are custom landing pages, static pages on the site, or a combination of both, in almost all circumstances email allows you to personalize the on-site targets.

This degree of personalization ran range from a dynamically-inserted greeting at the top of a page welcoming the visitor, to personalized product recommendations based on their previous behavior on the site.

Optimizing Direct Traffic

Visitors that arrive at your site from typing in its address or clicking on a bookmark are likely to be repeat visitors. They are also likely to enter your site through the home page.  Armed with this knowledge there’s a couple of things you can do to encourage these visitors to complete a website goal.

For visitors that you can positively identify as repeat customers, either by tracking cookies or a login, personalize their visit based on their preferences and past behavior. If they’ve previously made purchases, display items related to those purchases.

If they’ve visited the site before but exited before checking out, remind them of the items that remain in their shopping cart, possibly providing further incentives to complete the purchase by offering a discount or free shipping.

Optimizing for direct traffic at Zappos.com home page.

In the image above, the relatively generic block on the left is displayed by default. The block at right is displayed to a logged in Zappos user, and shows items related to recently-made purchases.

This sort of personalization may not be feasible, either because your site is not collecting data that can be used to dynamically modify page content, or because such mechanisms are beyond your technical capabilities or budget.

You can at least ensure that you keep your home page fresh, displaying new offers without interfering with the main navigation and search functions.  Return visitors will have something new to explore, and fresh content certainly won’t negatively impact the experience of first-time visitors.

Display, Social Media, Partnerships & Beyond

What I’ve outlined above are only examples of how site locations can be aligned with traffic sources:  whatever the source of traffic to your site, the more consideration you give to the source of traffic, the better your chance of converting those visitors when they land on your site.

Landing pages for a display campaign should reflect not only the message, but also the look and feel of the banner responsible for click-throughs to the site. The landing page for a pay-per-click ad will rarely work well for a campaign that originates from your Facebook page. Inbound traffic from a run-of-site campaign on a partner site will convert at a higher rate if the landing page is designed to leverage the interests or demographic profile of visitors to that partner site. And so on.

As a general rule, improving conversion rates based on specific traffic sources is a matter of fulfilling a visitor’s expectations. Your site will perform best when you try to determine, for each source, what promise motivated a visitor to come to your site, and then deliver on that promise once they arrive.

Thankfully, one of the easiest website elements to test is a button. Technically, buttons are usually discreet image files that can be swapped out with hardly any code modification required. In terms of procedure, there’s perhaps no more straightforward A/B test that can be run than measuring the comparative performance of two buttons.

Testing always plays a pivotal role in conversion rate improvement, but it is particularly important in crafting a winning button. User interaction with buttons is notoriously idiosyncratic:  the precise reason why a visitor will choose to click one button over a virtually identical alternative is next to impossible to decipher. The best-performing button is as often as not a surprise to the tester. And when it comes to buttons, there’s many things that can be tested.

The Button Call-To-Action

At a basic level, it is important that a visitor knows the consequences of clicking a button, and feels reassured that clicking will not result in any negative or unwanted consequences. It must at once inform and inspire confidence.

Clear messaging: the Firefox download button lets visitors know exactly what they're getting

As calls-to-action go, button wording like “click here,” “go” and “submit” tell visitors little about what clicking accomplishes (one study found that “submit” may be the worst way of enticing a user to submit a form).

This is particularly true of a ubiquitous call-to-action button that appears on all pages of a site, where the context of the button might not provide a visitor with enough information to want to proceed.

Make it obvious what action (and benefit) a visitor can expect from a button click.

This is not to say that a button’s call-to-action should be wordy or ingenious. Twitter was observed by multiple sources to be testing different variations on their sign up button.

One can only surmise that the extremely pedestrian “Sign up” that currently is displayed on their site performed better than livelier constructions such as “Give it a try” and “Let me in.”  Sometimes the best wording for a button is the simplest.

In situations where users might fear that their personal information will be shared, their computer might be at risk or that they might incur a charge, the call-to-action button (or text in close proximity to it) should be employed to reassure visitors.

A call-to-action button for a web-based software service might inform visitors that no installation is required. The submit button for an opt-in form where an email address is required might do well to include the message “no spam!” in big friendly letters.

An A/B test conducted for a solar energy company demonstrates how both clarity and reassurance in a button call-to-action impact conversions. Their original button that encouraged a visitor to “Click for Your Solar Consultation” was both vague (what’s a “solar consulation”?) and failed to remind users that clicking did not entail a commitment. Changing this text to “Get Your No Obligation Quote Now!” resulted in a 74% boost in conversions.

Instilling a sense of urgency has also been shown to lift conversion rates. On one hand, the simple addition of words like “now” and “today” (“download now!”, “sign up today!”) can improve conversion rates without making any particular claim about the timeliness of an offer. On the other hand, specifically stating or implying that an opportunity is limited might encourage a user to click a button for fear that they’ll be missing out by not doing so.

A Wider Funnel case study cites a 106% increase in bookings for a recreational vehicle company based on changing a persistent call-to-action to include a reminder that availability was limited, and changing the button text from “Go”  to “Get RV Rental Pricing and Availability Now!”

Perhaps obviously, if you’re offering something to a visitor without charge it is often beneficial to remind them that they’re getting something for free. A Visual Website Optimizer case study saw a 28% conversion rate lift for online address book Soocial simply by adding the words “It’s free” next a button with the call-to-action “Sign up now!” And, Firefox saw a lift from 9.73% to 10.07% for downloads by changing their button copy from “Try Firefox 3″ to “Download Now – Free.”

Like almost anything having to do with button performance, there are exceptions even to the seemingly logical proposition that a call-to-action that includes “free” should outperform one that doesn’t.

Successfully instilling a sense of urgency can be incredibly nuanced as well:  membership portal site Kajabi observed a 2% lift when they changed their button text to read “Get instant access now,” but a 252% increase in conversions when this was changed to read “Get started today.”

All of this to stress the importance of testing button call-to-action copy. However, much prior experience and common sense may suggest that one call-to-action is superior to another, it’s impossible to say which of two buttons will perform better without testing.

Button Prominence & Placement

Prominence: Dropbox employs whitespace in order to make their download button stand out

An effective button should be clearly visible on a page, and at least relatively prominent in relation to other elements.

On a landing page – and indeed for most pages, where this is feasible – a button should appear above the fold. The likelihood that a button will be clicked is greatly diminished if a visitor has to scroll to see the it.

For this reason, it is also common to see a call-to-action button placed both above the fold and at the bottom of a long page. Just a user is less likely to scroll down to click a button, they’re also less likely to go back up if they have scrolled past a button.

In a relative sense, button size matters as well. There shouldn’t be any question in a visitor’s mind of where they are meant to click. In one split test, Eric Graham reported a sizable lift in conversion rate simply by increasing the size of a opt-in button.

As size is relative, however, this does not mean that you should simply build bigger buttons, but that they should stand out from other page elements. This can also be achieved in large part by judicious use of white space:  a button surrounded by white space will be much more prominent than one which is lost in a sea of text and graphics.

Competing buttons: there's no doubt which is the more important button on this Amazon page

Proximity to other page elements is important as well. Obviously for an e-commerce site an “add to cart” button that’s right next to a product should perform better than one that’s further removed.

In other situations, it is important to keep a call-to-action button close to such things as value propositions, testimonials and feature lists that are intended to stimulate conversions.

Inevitably, a primary call-to-action button will be sharing real estate with other, competing buttons. The most important button should draw maximum attention to itself, though without interfering with the usability of buttons that have other functions (buttons must be in some way necessary, or they shouldn’t be appearing on the page).

Giving prominence to the most important button can be achieved by manipulating the size, color and contrast of this button.

Button Color & Contrast

Using contrasting colors: even in this thumbnail view, the Twitter "Sign up" button is unmistakable

Color has a demonstrable influence on whether or not a user will click on a call-to-action button, but determining precisely which button color will result in the highest number of clicks is, yet again, a matter of testing.

It has sometimes been said that a red buttons (and red text links) perform best:  by changing the color of their call-to-action button from green to red, Performable increased conversions by 21%.

Rightfully, Performable did not ascribe this to any inherent value of red over green, but speculated that it could be due to the color preferences of their specific audience, or how the red stood out in contrast to their website’s green color scheme.

Using contrasting colors is certainly a key factor in making a button stand out, particularly if there are many other interactive elements on the page. So it may be that increases observed in conversion rate from changing the color of a button may have less to do with the specific button color than how this color stands out on the website.

Certainly, a red button on a website full of red buttons is unlikely to be distinguished from its neighbors, and receive any greater attention as a result of its color.

Button Behavior

Button behavior: when mousing over this Zappos button, the pointer changes appearance and the button (subtly) changes color

It should be clear to visitors that a button is really a button – something you can click on that will result in a specific action – rather than simply a graphical element on a page.

Graphically speaking, there are a number of ways to do this, including embossing the button, placing the call-to-action text in a discreet bordered area, or offsetting the button from other graphical elements.

The button must also behave like a button when a user mouses over it. If a visitor hovers over a button and the cursor does not change from a pointer to a hand symbol, then that visitor is less likely to click.

For buttons that are not hyperlinked (and so do not automatically generate a hand symbol in the mouse over state) this can readily be accomplished with CSS.

A change in the button’s appearance itself on mouse over, such as a change in color, is a further signal (if not actually a small inducement) to the visitor that the button is clickable.

A Little Button SEO

In many situations, such as buttons used for the shopping cart functions of an ecommerce site, the button itself is not a hyperlinked element, and as such has no specific association with a target URL.

In many instances, however, a button is directly linked to a page indexed by the search engines. Adding an <img> alt attribute will provide the search engines with text they will associate with the target page:  if you are targeting keywords on that target place, you should employ them in your <img> alt.

Furthermore, an <img> alt attribute (or, depending on the browser, an <a> title attribute) may be displayed to a visitor when they mouse over the button, providing yet another opportunity to reinforce or augment your call-to-action (“start your no obligation trial!”).

What can the motivated marketer do to improve conversions today?

While testing will always provide you with the best guidance in making website conversion improvements, enough testing has been done collectively that certain conversion-facing improvements can be made in the absence of test data.

While you will always want to measure the impact of page changes on conversion rates by comparing data before and after those changes have been made, there are a number of measures you can undertake with relative safety that are almost certain to improve how many visitors complete a website goal.

1. Remove One Unnecessary Web Page Element

Singularly of purpose is important in maximizing conversion rates. In general, the fewer elements you have competing for a visitor’s attention, the more likely that visitor is to follow through on actions that ultimately lead a conversion, whether that conversion takes the form of an online purchase, successfully completing a form, or making a phone call.

If you’re only going to remove one thing to improve your conversion rate, the biggest bang for your buck is almost certainly going to come from removing an unnecessary form element. People don’t like filling in forms, and they also don’t like divulging personal information. By removing an unnecessary form field, you’ll increase the likelihood that a visitor will fill out and submit the form.

Review each form field and ask yourself two questions:  is it helpful to collect this piece of data, and is it necessary?

Gender declaration is a useful example here. If you ask visitors to select “male” or “female” on a form, what use do you make of that information? This is frequently the sort of data that’s collected solely to be able to report on visitor demographics, or is being collected for “future reference.”  In both of these cases, there’s no actual benefit in collecting gender information, so it should be dropped from the form.

Perhaps that information is used – say in differentiating subsequent email offers by gender. Then you need to ask if that information is necessary, even if it is helpful.

Put another way, is this essential information, or are you willing to see it omitted if this results in more conversions? If conversion trumps data collection, then at the very least make that a non-mandatory field, which will also likely improve your conversion rate on that form.

For sites where the conversion does not involve a form, or if you’re already asking for the bare minimum of information in your form, look at other places where an unnecessary page element can be removed. When a website visitor enters into the conversion funnel, you don’t want that visitor to be distracted from completing the goal at the end of that funnel.

The appearance of a secondary call to action, a link that unnecessarily provides a path out of the funnel, and even non-clickable visual elements that draw the visitor’s focus from completing an action can all potentially lower conversion rates.

2. Reduce The Load Time Of One Page

The longer a page takes to load, the less likely a visitor is going to stick around. According to a recent study, a one second delay in page load time can result in a 7% reduction in conversions. Furthermore, users are less likely to return to a website with pages that load slowly, and page speed is now acknowledged by Google to be a factor in rankings. There is simply no downside to making a page load faster.

Google has made diagnosing page load time easier of late with the introduction of page speed tools, including a simple form where you can enter a URL and page speed improvements are suggested. Common sense questions can augment tool findings.

Are any images being served that are being resized down with HTML that could simply be replaced with a smaller image? Are unnecessary scripts being run on the pages, such as tracking pixels that are no longer used?

Improving the load time of a single page may not send your conversions through the roof, but it’s a good place to start. However, as most sites are based on page templates, you may find that making one change has a positive impact on page load times across your website.

3. Add One Call To Action To Your Thank You Page

Thank you pages – the page that appears after a visitor has successfully ordered an item or submitted a form – are often missed opportunities. Yes, the visitor has already successfully completed the goal in this visit, but you want to facilitate future conversions by providing a follow-up action for that visitor to pursue.

Some of the calls to action you can consider adding to a thank you page include:

• Signing up for a newsletter
• Becoming a follower on Twitter or a fan on Facebook
• Inviting the visitor to share the offer they have just completed with others
• Providing a coupon for a future visit

Almost any call to action on a thank you page is better than simply saying “thanks” without providing any links for the visitor to follow. A visitor in this case is really left with only one easy course of action:  closing the browser window.

4.  Replace One Block Of Text With Bullet Points

Identify a page in, or closely related to, the conversion funnel that contains a large paragraph of text, and then rework that paragraph as a bulleted list. The bigger the text block and more convoluted the copy, the more likely it is that it is having a negative impact on conversions.

Among the reasons why bullet points may be more effective than text blocks:

• Bullet points are more effective for listing the benefits of a product or service, since each point is distinctly enumerated
• Special formatting stands out that much more in bullet points compared to a paragraph
• People find it easier to scan bullet points than a big block of text, which is important because the average visitor spends very little time on each page

The very exercise of reworking a wordy paragraph as a bulleted list can be beneficial. Self-promotional copy without useful information for the visitor doesn’t lend itself well to a bulleted list, and should be replaced with better information or deleted altogether.

There are any number of candidates you can look at for making this improvement, such as a product page with a verbose product description, or a wordy return policy FAQ. The existence or absence of a bulleted list won’t make or break your site’s conversion rate, but you’ll rarely be doing your website visitors a disservice by making the site’s copy easier to read and digest.

5. Use Analytics To Uncover One Actionable Insight

So you’ve got a fast site with nothing to distract or impede visitors in their journey through the conversion funnel. You edited your content and its presentation for maximum ease of readability. You’ve carefully considered possible post-purchase visitor actions and incorporated these into your thank you page.

Is there nothing you can do improve your conversion rate without testing?

Unless you’re not using analytics on your site (in which case it’s laudable, but bizarre, that you’re even investigating conversion improvements) the answer is almost certainly yes.

Sifting through your analytics with an eye to improving conversions, even without a specific purpose in mind, is almost certain to uncover at least one actionable insight.

Some possible lines of inquiry include:

• What page has the highest bounce rate on your site?
• Does the page deliver the promise of its primary traffic sources, or are users’ expectations being thwarted?  If so, how could those expectations be better met?
• Does the page vary substantially from pages with lower bounce rates, and if so, what aspects of the low bounce rate pages be carried over?
• Which page in the purchase funnel has the highest proportion of exits?
• Why might this be the case, and what improvements might you make to decrease the fallout rate?
• Of the top 25 or 50 keywords entered in site search, which has the lowest conversion rate?
• Does an existing page need to be modified, or a new one created, to better satisfy a visitor’s expectation for that search term?
• For ecommerce sites, is the site search for that low-performing query returning the right products, or does site search need to be tweaked to produce better results for that query?

Your site’s analytics are a mine of information that can help you improve conversions, but all too often analytics data is relegated to use in reporting or ignored altogether. Taking the time to perform some basic analysis on page performance will almost always result in at least one insight that will improve your site’s bottom line.

Are you targeting keywords that will bring not only traffic, but converting traffic, to your site? Is your page meta data optimized not only for high rankings, but for high clickthrough rates when your site’s snippets do appear highly in organic search results? How can the target page for any given keyword be manipulated to improve conversions from organic search?

Data from paid search campaigns can help answer all these questions.  PPC data can be invaluable both in initial SEO planning and for optimizing existing pages for improved conversion, largely because paid search can provide information on keywords for which a site does not currently rank.

Keyword Targeting

One of the most difficult things to assess in organic search is which keywords to target for optimal conversion. Or, more precisely, trying to determine which keywords will deliver a reasonable amount of traffic at a reasonably high conversion rate.

All keyword tools provide an estimate of the number of searches a keyword is likely to receive in a given period, but this says nothing about the conversion potential of those queries.

On one hand, a presumably high-traffic keyword may not actually deliver large numbers of visitors, because your site is a poor match for the intent behind the query. If you were to successfully optimize for that keyword, your site would appear in many search results, but few users would click through to your site from them.

On the other hand, a high-traffic keyword may actually drive large numbers of visitors to your site, but upon arrival those visitors fail to purchase an item, fill out a form, signup or otherwise complete a website goal. If you were successful in your optimization efforts for that keyword, you would drive large numbers of visitors to your site, but few users would end up converting (and as a result would be less likely to return).

Paid search can help define or refine your organic keyword targeting by providing data on both issues. The clickthrough rate on paid ad impressions can help you determine which keywords have a good organic traffic potential. This in itself can be extremely helpful for informational sites where ingestion of a single page is a valid conversion goal, and where other engagement metrics like time on site or pages per visit may not come into play.

More obviously, the actual conversion rate of paid search keywords can provide an important data-based clue for organic keyword targeting. At a page level, trying to decide which of a number of topically similar, but semantically different keywords to target is made immeasurably easier when PPC data is available: focus on what paid search has revealed as the highest-converting keywords.

While this sounds simple in theory, it may be more difficult to accomplish in practice, particularly in the not uncommon situation where SEO is managed in-house, but PPC by an agency. And a certain amount of guess work and extrapolation is going to be involved in prying search terms from broad and phrase matched keywords in different ad groups. But this is effort well taken. Otherwise, you may only discover after months of effort that your ranking success for “blue widgets” hasn’t resulted in the sale of many blue widgets.

Even in the absence of a paid search program, paid search in the form of estimated cost-per-click (as reported by the Google AdWords Keyword Tool) may help inform your organic keyword targeting. It stands to reason that if an advertiser is willing to pay $2.00 for “widget killer” but only $1.00 for “widget zapper” then it is likely that the former has been demonstrated to convert better.

Estimated CPC is an extremely blunt instrument and should be used with caution for SEO, but it can be helpful when trying to set optimization priorities among a number of similar, high-traffic keywords.

The Perfect Snippet

The snippet for a page in your site in the search engine results pages is comprised of the linked page <title>, its URL and, in most cases, the <meta> description for that page. This is roughly analogous to the elements of a PPC ad:  the linked ad headline, the display URL, and the ad text.

A knowledge of which ads have resulted in the highest conversion rate for a keyword or ad group can help you craft meta data that will be more successful in driving clickthroughs from search engine results pages to your site.

While the longest allowable Google AdWords ad headline is shorter than the maximum title tag that will appear in Google without an ellipses (25 characters versus 70), and the ad text shorter than a fully-displayed meta description (70 characters versus 156), successful ads can give you an excellent idea of what sort of copy resonates with searchers. Using ad copy to help fine-tune page titles and descriptions is especially helpful because there’s no straightforward way of testing the effectiveness of different snippets in organic search.

In some situations, successful ad copy can also be leveraged to craft messages on the page that will result in higher conversions from organic search. While a successful ad headline might not be exactly appropriate for a page title tag, it might be an excellent candidate for the on-page title or subtitle. Similarly, the messaging in the PPC ad might be used to improve the wording of a call-to-action on a page.

Leveraging PPC Landing Pages For SEO

In the happy event that your company uses landing pages as paid search targets, you can use these landing pages to help build permanent pages that will do a better job of converting organic search traffic, and may even improve the ability of those pages to rank for their target keywords.

This is really an extension carrying over snippet messaging to your organic target page to improve conversions, but on a bigger scale. Aside from using ad messaging that’s proven successful in paid search, you can also carry over other aspects of a PPC landing page that have proven successful through testing, such as the content of text blocks and visual page design.

The beauty of using PPC landing pages to improve your organic search performance – rather than landing pages in general – is that there is a relationship between query keywords and the effectiveness of the landing page in paid search that can be carried over in the organic realm. Where the traffic source for a landing page is not search, such as display advertising or a television commercial, then the same parallels may not exist.

Manipulating your site pages in this manner may not, of course, be possible. A PPC landing page for an ad with the headline “Buy 2 Leather Chairs, Get 1 Free” may not be altogether helpful in manipulating your standing ecommerce category page for leather chairs, even though both target the keyword “leather chairs.” Even in this situation, however, lessons derived from testing elements of PPC landing pages aside from messaging can help improve the organic search performance of indexed site pages.

Google AdWords Quality Score data from PPC landing pages can also be brought into play when trying to improve the performance of a page optimized for similar keywords. Landing pages with a high Quality Score are likely to better models for an organic search target page than those with a lower score, as Quality Score takes into landing page quality into consideration. Information from the AdWords keyword diagnosis report can also be helpful in building better organic search target pages.

Whether you are using data from PPC landing pages and paid search ads to improve on-page SEO or better define organic search keyword targets, paid search can be a great help in your SEO efforts. The biggest challenge to using paid search data effectively for SEO may be uncovering that data in the first place, so don’t be shy in approaching your in-house or external provider of PPC services for the information that can be used to improve the performance of your optimization efforts.

Like the Tweet button, the Like button is easy to install, and is highly standardized. Furthermore, the Like button is tied directly to Facebook’s analytics platform, Insights, which can provide you with valuable information about each Like button’s performance.

Display Type

The Like button is available in three formats. Standard, which shows the name of a visitor’s friend that has liked the page (if any friends have), and the total number of Likes; button count, which only displays the number of Likes in a compact format; and box count, that shows the number of Likes in a large block above the like icon.

Additionally, photos of the visitor’s friends who have also Liked the resource may be displayed.

Different Facebook Button Layouts and Variations

The same considerations that apply to the Tweet button are at play here, although the Like count is always displayed, so less active websites will probably not want to highlight the small number of Likes by using the box count option or displaying faces.

A light and dark color scheme are available, as well as different fonts; optimizing these may take the form of either closely matching your site’s look and feel, or perhaps trying to improve conversions by using a color scheme and font combination that make the like button stand out.

Positioning

The placement options that apply to the Tweet button are applicable here as well. Ensure that the Like button is placed where users are most likely to click on it in the context of the content, and don’t clutter up short pages with multiple like buttons.

Like vs. Recommend

Facebook provides two label options for the Like button:  “Like” and “Recommend.”

That recommendation  implies standing more firmly behind something than simply liking it does may discourage some visitors from clicking on a “recommend” button, but it may be the most appropriate verb for some types of content (and may increase other conversions associated with that resource, because it is stronger than “like”).  This is definitely something worth testing to maximize the number of likes a page receives.

Liking (& Commenting) vs. Sharing On Facebook

Prior to the introduction of the Like button in April 2010, visitors to a site could post both a link to a comment to their Facebook profile wall by using the Facebook share button (introduced in October 2010).

Some users have continued to employ both the share and Like buttons, both because of the commenting functionality of the share button, and the greater visibility granted to shared items on Facebook.

However, as a result of a February 2011 update to the Like button, Liked items now receive the same visibility on Facebook as shared items, and users are given the option to add a comment to a link once the Like button has been pressed.

For these reasons, and the fact that Facebook is no longer developing the share button, there is no reason to carry both buttons.

From a conversion perspective, the presence of two Facebook buttons with different labels that accomplish basically the same thing is not helpful;  if you’re carrying the share button, consider retiring it and using the Like button exclusively.

Read on for other useful tips in this series:

• How To Convert Visitors To Tweeters
• How To Convert Website Visitors Into Twitter Followers
  • Plus: How To Use “Direct Following”
• How To Convert Website Visitors To Facebook Likes
• How To Convert Website Visitors Into Facebook Fans
• 7 Approaches To Engagement Conversion On Twitter & Facebook

This is because the following action takes place on Twitter itself, rather than on your website:  you’re essentially encouraging a user to click on a link. As such, there’s more optimization and testing opportunities.

Call-to-Action

Your call-to-action can take the form of a text link, a linked graphic or both. This gives you the opportunity to create and test any number of calls-to-action until you uncover the formula that delivers the highest number of followers.

Best practices for other calls-to-action can be called upon here:  the use of action verbs, brief but compelling text that provides an incentive to click the button, and so on.

Positioning

Factors to consider here are the closeness to the Tweet button, the positioning of other network join buttons (especially your Facebook call-to-action) and the relative effectiveness (or ineffectiveness) of a call-to-action to follow in a place like the sidebar.

Tweet Stream

A stream of your recent tweets may provide an additional incentive for visitors to follow you if, of course, they like what they see.  This technique will obviously work best if you have a very active Twitter account that displays a high level of interaction.

Direct Following With Twitter

It is technically possible to have a button that allows users to follow you without leaving your site.

To accomplish this with any degree of customization entails creating and registering an @Anywhere application, so is really only suitable for those with development skills.

Tweetmeme also provides a follow button, but there are few customization options available. These options are worth exploring if you have the skills to implement them and want to try and decrease fallouts in the following process, but users are fairly well-accustomed to initiating a follow on the Twitter account page itself.

In any case the official Twitter button itself provides a mechanism for following without leaving your website. Any visitors that successfully complete a Tweet from your webpage are subsequently presented with the option of following one or two Twitter accounts.

 

Tweet Button Follow Recommendations (Second Recommendation w/ Custom Description)

This, in itself, is another compelling reason to use the official Tweet button, as this functionality is built in.

That you can suggest two Twitter accounts to follow in the pop-up is particularly useful for content sites, where often both the author of a post or article and the domain itself have Twitter profiles (accounts the visitor is already following do not appear).

A line of text can optionally be inserted above the profile block, providing at least a nominal opportunity to create and test a call-to-action.

Read on for other useful tips in this series:

• How To Convert Visitors To Tweeters
• How To Convert Website Visitors Into Twitter Followers
• Plus: How To Use “Direct Following”
• How To Convert Website Visitors To Facebook Likes
• How To Convert Website Visitors Into Facebook Fans
• 7 Approaches To Engagement Conversion On Twitter & Facebook

The alternate method of enticing visitors to become fans is simply to link to the Page on Facebook, where there is a prominent Like button at the top of the Page.

There are advantages and drawbacks to each method.

The obvious benefit of a direct Like is that the user is connected in one step, without the danger of failing to click on the Facebook Page Like button. The biggest drawback is not being able to control the call-to-action or its appearance: linking to a Facebook Page is equivalent to linking to a Twitter profile, and provides the opportunity to closely associate following on Twitter and becoming a fan on Facebook.

What optimization can be performed at the page level depends on which method is used, though of course both a direct Like button and a linked call-to-action can be used on the same page, and their performance compared.

Positioning

If using the Like button directly, it is important to separate this from any other Like buttons on the page, and ideally identify for the user what it is they are about to Like. When linking to a Facebook Page, the same positioning considerations apply as with a Twitter linked call-to-action.

Call-to-Action

For Page Like buttons, “Like” seems the most natural label, and “recommend” seems awkward. For linked Facebook Pages, you can do away with “Like” altogether, and try things like the older, descriptive “become a fan on Facebook” or even “follow us on Facebook.”

Like Box

The Like box is a sort of extended version of a Like button for Pages, and displays a “Find us on Facebook header,” the number of users that Like the Page, recent posts from the Page, pictures of profile photos and, of course, a like button (“recommend” is not available).

The header, post page stream and photos are all optional.

The Like box takes up a fair of amount of real estate with all of its functions enabled, but can be pared down and employed as a beefier call-to-action to Like a Page than is offered by the Like button itself.

For active sites with large numbers of users and good content, the Like box may serve as an enticement by showcasing the Page’s usefulness and popularity.

Read on for other useful tips in this series:

• How To Convert Visitors To Tweeters
• How To Convert Website Visitors Into Twitter Followers
  • Plus: How To Use “Direct Following”
• How To Convert Website Visitors To Facebook Likes
• How To Convert Website Visitors Into Facebook Fans
• 7 Approaches To Engagement Conversion On Twitter & Facebook

There are a number of things that can be tested in order to improve the number of Tweets a page receives.

Display Type

The Tweet button is available in three versions, all versions sharing the same color, size and label (“Tweet”). You can choose a button that does not display tweet counts, one that displays the tweet count next to the button horizontally, and one that displays the tweet count above the tweet button.

 

Tweet Button Styles

For sites where pages receive a lot of Tweets, the vertical option puts the highest emphasis on the Tweet count and is probably your best bet (because of the “bandwagon effect,” whereby something demonstrated to be desirable is more likely to be perceived as desirable).

For sites with little traffic, or things like product detail pages where a low Tweet count might potentially discourage tweeting (or give a product or resource the appearance of being “unpopular”) a button without a count might be the way to go.

Positioning

What works best:  a Tweet button at the top of the page?  The bottom of a page?  Multiple locations?

Obviously, for very long pages like in-depth articles, a bottom placement (or top-and-bottom placement) seems logical, as visitor may not return to the top of the page to click on a tweet button that they’ve long ago scrolled past.

For a page where all or most of the content is above the fold, such as a video or product page, multiple buttons may clutter up the page and dissuade the visitor from taking other actions. Another possibility for the technically-proficient is a fixed twitter button or share block, that stays in the same position on the page as a user scrolls.

Twitter Reactions

A stream of tweets being made about a page, or even tweets topically related to the page, may encourage visitors to click on the tweet button themselves in order to join in on the conversation (although for content-based sites the most obvious real estate to house Twitter reactions is usually taken up by comments).

Unlike the relative simplicity of adding a tweet button, a certain amount of heavy lifting is involved to include Twitter reactions:  manual scripting, a custom script or a third-party tool is required.

Read on for other tips in this series:

• How To Convert Visitors To Tweeters
• How To Convert Website Visitors Into Twitter Followers
  • Plus: How To Use “Direct Following”
• How To Convert Website Visitors To Facebook Likes
• How To Convert Website Visitors Into Facebook Fans
• 7 Approaches To Engagement Conversion On Twitter & Facebook