Tuesday, 29 May 2007

PPC & Usability - Some Tips

The greatest advantage PPC has over other forms of marketing is its measurability. It's possible to see, down to each individual click, what revenue has been earned for what amount spent. In fact, just about every metric imaginable is measurable - ad views ("impressions"), clicks, click-through-rate, total cost, average cost per click, revenue, actions, cost per order, cost per action, and so on. Couple this with web analytics packages, and you have the complete tracking solution. Compared to other media like, say, television, it's head and shoulders above the competition.

However, the caveat to this is that a PPC campaign is only as good as the results gained. And to get good results, you need to have a streamlined, water-tight campaign from beginning to end - that's from initial impression to final action. You need to make sure your initial searcher - pulled in by your PPC ad - turns into a confirmed purchaser. What I'm trying to say is, no matter how good your keyword choices, ad positions, creative messages and qualifiers, it counts for little if your destination web pages aren't usable. If they're not usable, you could have the greatest PPC campaign in the world and your ROI will still suck.

This is why usability is a PPC marketer's friend. The more usable the destination website, the more likely the PPC visitor is to convert into a customer (thus adding to your magic ROI number...or cost per order, cost per action, or whatever your chosen metric). Here are some basic tips on best practice in PPC usability:

1) Qualify visitors as much as possible before reaching your website - your keyword choice and advert creatives are your first opportunity to influence how usable the destination site will be. You can filter out unlikely customers and encourage "good 'uns" by choosing the right keywords and giving out the right messages in your copy (e.g. if you are a luxury brand, with what is in all likelihood a fancy, flash-based website, steer clear of the words "cheap" and "discount" and use words like "luxury" or "designer" in your copy). By properly qualifying your visitors, those arriving on your site are likely to have a more positive experience (which is ultimately what usability is all about).

2) Deep link wherever possible - don't assume your home page is always the best destination for your visitors. Whilst this is in all likelihood the most visually impressive site on your page, it is probably not closest to the conversion goal. A good rule of thumb is, the further down the conversion funnel you can send your visitors, the better (if your visitor is looking for pink golf balls, take them right to that product - not to the homepage, or even the main golf balls page).

3) Provide information - consumers like to be provided information and providing it on your site makes conversion more likely. So, if you only offer free delivery on orders over $10, let the visitor know this; if you don't deliver outside the UK, make this visible too. Removing doubt and providing info creates customers.

4) Keep things consistent - if your homepage has links down the right hand side, and drop down menus on the top, try and make sure these core navigational elements remain consistent across all your other pages on the site. Likewise, if you highlight links using underlined blue text on one page, don't use italicised red text on another. Try and keep things consistent. By doing this, users know exactly where they stand and will feel comfortable and confident about navigating through your site.

5) Create a clear sales funnel - know where you want your visitors to go when they land on your site, and make this pathway as clear and easy to navigate through as possible. For example, if you want people to send you a contact form, make the link to this form as obvious as possible (without it being so big it interferes with the basic usability of the site).

6) Make checkout nice and easy - the number one stage at which customers drop out of the sales funnel is during checkout. We've all done it - you get all your purchases ready in your basket, go and total up the costs, then, for whatever reason, decide not to proceed. There are two main reasons - one is that checkout is somehow complication (perhaps requiring registration first, or maybe it asks for too much information or is vague in what info is required) or there's something dubious or unexpected at the checkout point (hidden charges or an error message of some kind). In any of these cases, expect conversion rates to plummet. It really is in your interests to ensure checkout goes well - after all, this is the home straight to conversion, and the point to which all your hard marketing work has led!

Tuesday, 22 May 2007

Checkouts - Google vs Paypal

Since the launch of Google Checkout, there has been much talk about whether Paypal's dominance of the online checkout market would be under threat. It seems the launch of the search giant's offering has been taken seriously by Paypal, who have recently launched a new set of incentives to ensure banks and technology providers use their Express Checkout service rather than Google Checkout. Paypal are now offering a $2500 bonus, tech support and marketing materials to merchants who sign up to their Express platform.

I think Paypal are right to launch this counter-attack in the "Checkout Wars". After all, with very little to separate the two systems in terms of functionality and cost, Google have a subtle advantage in their favour - they have the Google Checkout badge to offer. A number of commentators have already noted the increase in CTR the presence of the logo brings PPC advertisers. It seems Google were very aware of this fact, and made a bold move early on to make the GC badge even more visible in the listings (see below).


The increase in clicks and quality score resulting from adding the GC badge to listings will undoubtedly lead some merchants to make the switch, if only for this reason. However, this move may yet prove short-sighted. After all, a lot of these additional clicks will arise from pure curiosity - they see the big flashy Google Checkout badge and want to see what it's all about. As such, I'd expect these additional visitors to convert at a lower rate than those who click regardless of the GC badge.

Nevertheless, Google Checkout is attracting interest, clicks and, more importantly, merchants. Whether it will prove a long-standing competitor for Paypal remains to be seen. It's certainly a pretty interesting battle though, and one Google will be desperate to win.

Sunday, 20 May 2007

Bidding on Branded Terms - Why You'll Lose Customers If You Don't

When creating a keyword bank for your PPC campaign, your first port of call should be your branded terms. Protecting your brand and ensuring you funnel all available searchers to your site should be priority number one. Why? Well, there are two main reasons:

1) brand terms convert up to ten times better than generic terms
2) the perception of your brand is weakened if you're absent from the sponsored links

Let's address point one first. Why do brand terms convert better than non-branded terms? Well, you've heard of the "buying cycle" right? The further you are down the buying cycle, the more ready you are to buy. When someone searches for a branded term on a search engine, they're indicating a strong intent to buy. When they're searching for a generic term, it indicates they're still in the researching period.

To illustrate this, let's consider Mr Jones, who's looking to buy pet insurance for his family's new puppy. When he first goes onto Google, he has only a broad idea of what he's looking for. He knows he wants to insure his pet, and he knows he wants to get a good price, but doesn't know too much beyond this. For this reason, he enters "pet insurance" into Google. Now, this is about as generic a term as you can get within the pet insurance market. The SERP that comes up will likely a broad spectrum of insurance companies, some specialising in pet insurance, some not, some offering insurance on all types of pets, some not, some cheap, some not, and so on. Mr Jones will then think to himself, "well, it's a dog I want to insure, so let's try narrowing things down that way". He then enters "pet dog insurance" into the search engine. Now his results are a bit more targeted. The engine shows him only results for canine insurance. He looks at a few sites from the listings but decides they're too expensive. He then enters his last generic search - "cheap dog insurance". This brings up some great sites, with just the deals he's looking for. One in particular catches his eye, being offered by BarkinglyCheapInsurance.com. He jots down the details, goes and chats with his wife, has his dinner, then sits back at the PC later that night. He can't remember the url of the site he wants, but enters "barking insurance" into the search engine. He finds the site and makes the purchase. This decision process can also be initiated offline - if Mr Jones saw a TV ad for this great insurance deal, or perhaps was recommended by a friend. It's all the same - ultimately what happens is he comes back and looks for your brand in order to buy.

Now, suppose he hadn't been able to find BarkinglyCheapInsurance.com when he went and looked after dinner. Perhaps the site didn't rank well organically for its brand search, or Mr Jones had misspelled the brand and therefore the engine didn't serve his listing. That site stands to lose a customer. However, I hear you thinking - "well, my site ranks well for my brand - number one in fact - so he would have no trouble finding it". Well, this maybe so, but suppose he'd searched for your brand and EvenCheaperDogInsurance.com came up in the sponsored links with an offer that seemed even better (he'd looked for you at the top of the sponsored links, but you weren't there). He'd most likely click on this better offer, and you'd have lost that customer....even though he was searching for your brand!

Now, I can even hear some of you saying "well, we're number one organically and have our brand trademarked, so nobody else can show an ad for our brand". You think this covers you? It almost does. Most people will find your site - around 75% in fact, with the other 25% either not seeing your listing or deciding they've seen another listing that actually looks better. However, if we look beyond merely gaining that brand click, we'll see that your conversion rate actually suffers if do not show a PPC ad. A number of studies have shown this, but my favourite is by optimizeandprophesize.com. The reason your conversion rate improves when you present a searcher with both an organic and a PPC ad, is that the perception of your brand is strengthened and you also have the ability to send the searcher a powerful call-to-action marketing message within your PPC ad. As a marketer, you have far less control over the message displayed in your organic listing. With PPC, you can tailor it however you like. The net result of this is that you convert more searchers into customers.

There's no denying that generic terms are a very important piece of the PPC jigsaw. They're what drives the high volumes of traffic, and are what places your brand in the front of your searchers' minds. However, you must then ensure you close deal. The best way to do this is by maintaining a solid and persuasive presence in the SERPs when someone searches for your brand. If you can, get other advertisers removed from your brand using a Google Trademark Claim (see Google's help pages on how to do this). However, even if you can, it's vitally important that you gain great presence in both organic and paid results for your brand. By doing this you ensure your brand will be found, you ensure your brand looks strong and credible, and you ensure your searcher reads the kind of positive, call-to-action message required to turn them into a customer. All this will ensure you get your sale.

Oh, and did I mention brand clicks are cheap? If you run your campaign right, you should be able to get brand clicks for under 5p. Now, if this doesn't finally convince you that brand bidding is essential, I don't know what will!

Wednesday, 16 May 2007

Gambling & PPC - The Lowdown

With the revelation that Yahoo! will be following Google's lead and banning advertisers from promoting online gambling sites via PPC, it seems a good time to take a look at the current situation with regards to this subject, and what the implications are for advertisers.........

Back in the day, when PPC listings were full of sex, drugs and rock and roll, it was ok to advertising gambling sites and have outwardly promotional gambling messages in creatives. After all, these were more often than not shown when someone was looking for an online gambling site, so advertisers were simply doing what comes naturally - meeting consumer needs. However, in 2004, Google announced it would no longer allow advertisers to promote gambling sites or services through PPC ads. This obviously hurt the online gaming industry, leaving many sites struggling to bring in new visitors, since their most potent channel was now rendered impotent.

Yahoo! have now, unsurprisingly, followed The Big G's lead, and launched their own ban on gambling ads. MSN are now the only major search engine in the UK still permitting gambling-related PPC ads. This move by Yahoo (and previously Google) has been met with a mixture of approval and anger, the latter very much the feeling in the online gaming community. A number of commentators have highlighted the fact that pornography suppliers, such as Max Hardcore, can freely utilise the PPC medium (try searching for "max hardcore" on Google).

The feeling of injustice has been intensified by the launch of the Gambling Act 2005 in the UK, which permits advertisers to promote their businesses on television. This means gambling sites can advertise freely on television, the largest reach medium around, to an untargeted and multi-faceted viewing group, but cannot place targeted ads online to help customers find them. This seems a quite farcical situation.

However, whilst advertisers are not permitted to promote gambling in their adverts, they can still bid on gambling related terms and have ads that do not directly promote gambling with actual cash. Carry out a search for "online gambling" on Google and you'll get a raft of ads. However, you'll notice that almost all the ads present include the words "free", and those that don't do not mention money. What's more, click through an ad and you will be taken to a site where no actual gambling takes place. It's all free games and "fun" activities.

However, despite the seemingly innocent ads and sites here, there is a serious marketing effort at work. After all, where's the profit in free games? No - what these sites do is bring gamers to their sites through gambling related phrases, using non-gambling related ads, to play fictional gambling games; but before they can play the game, they must enter an email address. This is where the marketing will come into force. The site, as soon as the email address is gathered, will begin an aggressive and, often, highly targeted email marketing campaign to bring the gamer back to their parent site to play for real money. This is how they get their new customers via PPC. It's not quite as direct as most PPC adverisers, but they get there in the end. The conversion rates for this kind of email follow up is high, too, so it's not all bad for these sites.

Nevertheless, it's my opinion that, particularly in the light of the current relaxed regulations surrounding gambling advertising on tv, the ban should be lifted from PPC. Especially since the ban is not a complete one anyway - if Google were really serious about stopping sites getting new gamblers via paid ads, they'd stop advertisers bidding on the phrases. It's a no-brainer - if you don't allow people to bid on gambling-related keywords, then they can't carry out directly gambling-related PPC activity. Instead, Google and Yahoo have brought in half-measures that, although very damaging to sites' PPC efforts, do not go the whole way to banning gambling advertising through PPC.

I'm all for promoting responsible gambling behaviour - gambling addictions are a serious matter, and should be treated as such. However, I think it's far more dangerous for a vulnerable person to be exposed to adverts via a high-reach, low-targeted medium such as television, than have the risk of finding a PPC ad for gambling after typing in "online betting" in a search engine. After all, if that's what they're looking for, they can just click on an organic link anyway..........

Tuesday, 15 May 2007

Keyword Research - The Building Blocks of your PPC Campaign

The skill and expertise involved in creating and managing an efficient paid search campaign are often overlooked. It takes quite a unique blend of talents to cope with the number-crunching, copy-writing, in-depth analysis and general stress levels associated with PPC.

One of the most undervalued and under-appreciated tasks in the whole PPC process is keyword research. There are literally thousands of Adwords accounts out there that contain no more than a handful of keywords, and yet the advertisers are surprised when CPCs and high and ROI low. I've seen accounts that contain a single ad group with just a dozen keywords - and yet they are advertising six or seven different products. This advertiser was appearing for around 70% of relevant search phrases, which on the surface seems ok - however, when you consider that their CPCs were inflated almost three times above those of their competitors, and the CTRs were around half the industry average, the failings become more clear.

At the outset of a campaign, creating a comprehensive keyword list, organised into small, relevantly themed ad groups is crucial to good future performance. Including keyword variations, plurals and misspellings will help you reach even more of your target audience and, because these "secondary" phrases are less competitive than your more obvious, core keywords, you'll get clicks at a cheaper cost. Also, because secondary phrases tend to be more specific than your primary, core phrases, they'll convert better, too.

As an example, consider a credit card company. Primary, core phrases for this campaign would be "credit cards", "visa card" and "apply for a credit card". These phrases are big, big traffic drivers, but are also competitive. The volume of traffic they drive is partly responsible for the high level of competition, but it's also the fact that these phrases are very "obvious". If you ask a non-savvy internet user what you might type in to get a credit card, they'd likely tell you "credit cards" or "apply for a credit card". As such, every advertiser in the market is going to have these phrases in their accounts. And, due to the auction nature of the PPC system, prices go up as the number of advertisers goes up. You're likely to pay £7-8 ($12-13) for a click on the term "credit cards" and £6-7 ($11-12)for the term "apply for credit card". With CPCs this high, you need to be converting visits into leads at a pretty high rate to keep your campaign profitable.

At the other end of the keyword spectrum you have less competitive, less generic keyword that, although driving less volume, are more targeted and cheaper to appear for. By including a high number of these "keyword tail" phrases, you can drive quite substantial volumes of high quality traffic at a fraction of the cost involved with more generic terms. In the example above, longer tail phrases like "apply for a credit card online" or "get best deal on credit cards" will have far lower CPCs than generic terms, and will likely convert better due to their more targeted nature. Throw some branded terms in there - both your own brand and, if possible, some competitor brands - and you'll bring your average CPC for the account right down, and your conversion rates should rise too.

Here are some useful tools for compiling your PPC keyword list:

Yahoo's Keyword Assistant tool (free)
Google Adwords' Keyword tool (free)
Wordtracker (fee)
Keyword Discovery (fee)

Yahoo's tool is very good.....when it works. It's lack of reliability is a major problem though - it will work at times, but other times you won't even be able to find the site. The figures are reliable, as they're based on actual Yahoo search volumes from previous months. Google's tool is reliable in terms of functionality, but its figures are highly dubious. We've seen results come in at 50% below actual figures, and this is not uncommon. Both Wordtracker and Keyword Discovery are good tools, but obviously you need to pay for those :s

The process of good keyword research is long and tedious, but done properly can provide the foundations for a rock-solid campaign. The key is to start with generic phrases (you'll always need these in your campaign for traffic volume, branding and general visibility purposes), and then expand the list out to cover all long tail variations. So, say I'm the credit card company above, I would start my research with the most obvious phrases ("credit cards", "visa cards" "apply credit card", "credit card online" etc) and then see what secondary phrases arise from these. By inputting these keywords into the tools listed above, hundreds of variations should come to the surface. These "seconary" phrases can then be categorised into targeted ad groups, in which you can place a highly targeted advert (so in this case you might have a "Credit Card - Online" campaign, inside which you'd have the ad groups "Apply", "Gold", "Silver", "Platinum", "Low Interest", "0% Interest" etc). Try and keep high volume, generic phrases within their own ad groups - this will help both with the management of these phrases and overall performance tracking of the campaign.

Take time over the above process. Explore all phrase variations, and don't forget your plurals and your misspellings. By the end of the research you should have a number of campaigns laid out, with themed ad groups full of a mix of high traffic, generic phrases and specific, highly targeted tail phrases. If you don't, go back and see what you've missed!

Sunday, 13 May 2007

The Rise of Malware - What Impact on PPC?

Google's recent report on malware - interestingly titled "The Ghost in the Browser" (I feel a movie script coming on....) - revealed that one in ten pages on the internet is capable of downloading malicious code onto user's PCs without their knowledge. This news, which follows fast on the heels of the recent Times Online reports on "Adwords hackers", has real implications - not only for the general internet-using public, but also those who make a living from advertising online.

PPC as a medium has always, among certain quarters, had an unfavourable reputation. In its early days, unfair association with such irritants as pop-ups, pop-unders and flashing banners meant many advertisers stayed away from paid search advertising, instead going for an SEO-only approach to being found online. During the dot com boom, when PPC really flourished, it was dogged by massive click fraud problems. Indeed, many will argue that the problem of click fraud remains major even today. The manner with which PPC providers - and search engines in general - came and went in the early days didn't exactly inspire confidence either (only GoTo.com really stayed the course in the early years....later becoming Overture, then Yahoo Search Marketing).

More recently, Google themselves have had their reputation tarnished by the revelation that their own PPC system - Google Adwords - had been hi-jacked by hackers intent on spreading malicious code. Ads associated with around 20 search terms had to be removed from the sponsored listings, where the destination website (disguised by a fraudulent display url, impersonating a legitimate business) attempted to install password-stealing software onto the user's PC (the software is known as 'key-logger', which records user keystrokes, and is ironically employed by many anti-criminal organisation such as the FBI and MI5).

The issue of whether or not advertisers should be allowed to use a display url vastly different to their own domain name is another issue, too large to be discussed in this post. However, the more general problem of an individual using PPC to carry out criminal activity cannot be ignored. At the end of the day it comes down to the providers - those making the money - to do something about the problem. More stringent review of new advertisers could be one way to do this, although the flip argument to this is that this will increase submission times for adverts and therefore annoy legitimate businesses looking to get PPC up and running. Perhaps making it easier, and quicker, for users to point out illegitimate ads to the engines would help (perhaps a beefed-up version of Google's blog spam reporting system might work well). A stronger partnership between anti-virus providers and search engines might also help - perhaps if there existed some kind of system that quickly scanned the destination site of a sponsored listings, before the user went there, situations like the one described above could be avoided.

Whatever the solution, one thing is certain - PPC providers must step up to the plate in the fight against malware, much like they have done recently with click fraud. Providers like Google have made great strides in eradicating click fraud; their click fraud report, released earlier this year, showed that fraudulent clicks accounted for less 10% of total clicks (most of these are filtered out by Google's systems, so will not be paid for by the advertiser). Yahoo made a similar announcement, stating just 12-15% of their clicks are invalid, shortly after; they then went one step further and appointed a VP of Marketplace Quality.

Providers need to show the same levels of conscientiousness with PPC malware, as they have with click fraud - not only to maintain quality SERPs, but also ensure the long-term future of PPC as an effective, ROI-driven advertising medium for businesses.