Conversion Rate Optimisation (CRO) describes the process of analysing the performance of your website/landing page or other marketing material, how people interact with it and making changes that will improve the numbers that convert into customers.

In this article I’d like to discuss one of the recent successes we’ve had with implementing CRO for one of our clients. For privacy reasons I can’t mention anything too specific but I hope that what I do include is enough for you to try CRO out yourself.

The Scenario

The client in question had tried out SEO a few years previously with little success, and after a period of time stopped all activity. Unsurprisingly the number of SEO conversions had not grown since and was not particularly high to begin with.

Alongside the normal link building and on-site optimisation that was carried out, we took an in depth look at the analytics data that had been collected over the last few years.

Identifying ‘quick win’ CRO opportunities

The first step was to create a non-branded organic segment in Google Analytics to remove visitors finding the site using a brand term – we know these visitors have prior knowledge of the client either from other marketing activity or from previous visits so are not the main focus for this particular exercise (although there are of course CRO opportunities there).

How you filter out brand terms is up to you – in this case we’ve used a RegExp filter to find keywords containing certain fragments and to remove that traffic from the report.

Poor Performing Landing Pages

With the non-branded organic segment selected we studied the conversion funnels starting at the SEO landing pages and ending with the confirmation page. The funnel conversion rate seemed to be fairly healthy, so the problem must be with the amount of people entering the funnel – in other words the number of people moving from the landing page to the second stage of the funnel.

Identify pages to analyse

With the non-branded organic segment selected, navigate to the top landing page report under content.

On the report that follows add a filter to display landing pages with a high bounce rate (see below).

You should be left with landing pages with a bound rate higher than 65% (or whichever percentage you decide) that people arrived at using non-branded keywords. We do this to focus our initial efforts on pages that are particularly poor performing, but you can reduce the bounce rate filter or remove it altogether to incorporate more pages.

Keywords Used

Finally select a landing page from the report that receives a reasonable amount of traffic. The CRO we’re carrying out in this instance would be wasted on a page that receives just 2 or 3 visits a month.  Anything receiving 50+ visits a month is worth looking at in these early stages – if the CRO we carry out is effective here we get very quick wins and can roll it out across all relevant but low traffic pages afterwards.

With the landing page selected, pull up the entrance keywords report and answer this simple question:

Are the keywords visitors use relevant to the topic of the landing page?

If the answer is yes, then carry on. If the answer is no you’ve got a separate problem!  Make a note of this so it can be addressed later then go back a step and select a new landing page.

In the case of this client we saw a very high bounce rate for the landing pages (typically 70%+), despite the search terms used to find the pages being very relevant – people were finding what they wanted but leaving without visiting another page on the website. In this case there’s clearly something wrong with the page!

The Landing Pages

In the extremely simplified image below is the layout of the landing pages in question. There was a call to action but it barely stood out from the rest of the page. Of course it wasn’t all black and white, but the effect was the same.

There’s an argument for a complete redesign but in this scenario we’re looking for quick wins without the need for major redesign/redevelopment of the website.

The Solution

The problem here is that visitors arrive at the website but are given no clear direction or suggestion on what to do next. While there is a call to action in the right hand sidebar, it’s lost between other elements.

Our suggestion was to place a call to action button under the introductory paragraph designed specifically to stand out against the rest of the page. The actual text on the button included ‘Free’ to further entice a visitor to click through.

By placing the C.T.A high on the page and making it clear both in terms of colour and wording, it was hoped that the conversion funnel would be fed.

You may want to take this a step further with your own pages – another call to action at the bottom of the page may be useful if the length causes users to scroll.

The Results

In the month after the C.T.A buttons were placed on the landing pages the following improvements were seen:

• Bounce rate decreased by 15%
• Conversion Rates increased by 72%
• Sales increased by 89.47%

Try it Yourself

• Create a non-branded organic segment*
• Browse to the ‘Top Landing Pages’ report under Content
• Add a filter to show landing pages with a bounce rate of 65%+ (or reduce as required).
• Check the keywords used to find the page for relevancy
• If relevant look at the landing page – is there a clear call to action?
  • Does it stand out from the rest of the page?
  • Is the text descriptive and enticing?
  • Implement the new CTA (whether form, button or whatever is applicable)
  • Analyse results over the next few days

There are many ways to improve the performance of your website and just as many to go about identifying them but I hope this scenario gives you a starting point to delve into your analytics and start fine tuning your website for conversions.

*You can apply this entire article to PPC (paid search) just by creating a non-branded PPC segment in Google Analytics, something I would suggest regardless as it gives you great insight into how you perform with people who are unaware of your brand and who often represent new rather than repeat customers.

Boasting free 5GB for its users (and opportunities to increase storage options) Drive gives you the opportunity to create and share new documents in real time on Mac, PC, Android phone or tablet.
Features include
• Post comments
• Receive notifications
• Image recognition
• OCR – Optical Character Recognition (recognising text in scanned documents)
• Smart tagging – tag items stored on your drive. For example, If you carry out a search, Drive can recognised uploaded images and present these in results

Available at drive.google.com, it works just the same as Dropbox. It creates a folder on your system (once you’ve downloaded the application) and even comes with an app for Android so you can share across multiple devices (Ipad & iPhone app to follow soon).

While this isn’t a revolutionary service, the fact that it carries the Google brand and integrates with other services many people are already using, could give it the head start it needs to seriously rival competitors such as Dropbox.

It’s worth noting that Dropbox’s CEO Drew Houston is reported to have rejected a nine figure (£100m+) offer from Apple for the company who soon after launched their own cloud storage solution; iCloud – a strong indication that the big players value this market very highly indeed.

Google Drive pricing starts at $2.49/month to upgrade from 5GB to 25GB, $4.99/month for 100GB and for those with greater storage needs, 1TB for $49.99/month.

What’s the latest Google invention? Their brand new prototype product from ‘Project Glass’ – Google Glasses

Google Glasses are a brand new head mounted display (HMD) giving the user hands free, voice activated access to the digital and comms world, whether it’s social platforms, Google maps, emails, texts or calls.

The prototype is a small screen in front of your eye that then overlays data such as directions, social media and communication notifications. Technology is said to include 3G or 4G, GPS and a built-in camera.

Check out the promotional video below.

There has been some confusion as to when (or if) this product is being released for purchase to the public. Whilst it has not been confirmed when we’ll be able to get our hands on one, The New York times reported in February that “the glasses will go on sale to the public by the end of the year and they are expected to cost around the price of current smart-phones, $250 to $600” however if these did become available, how practical would this new technology actually be for end users and for businesses?

For the end user, it gives them the chance to be continuously connected to friends and family. For businesses it creates new gateways for communicating with virtual teams and colleagues.

However there are a million and one questions that arise from these glasses, and whether it is actually a commercial benefit
• Will advertising be effective or even appreciated?
• What new behaviours will this new technology bring?
• How would people search on the internet with limited screen space for information?
• With this in mind, how will companies differentiate their USPs?
• As it’s designed for people who multitask (users are walking, eating, shopping etc) will they even be paying attention to display ads?

As the prototype begins its testing, it seems that only time will tell how we can all embrace this new exciting technology and how it will shape our modern day culture of mobile communication.

Looking after your ad words account can be challenging at times; CPC bids, negatives, ad copy changes, analytics analysis and strategies etc, and one challenge that is regular within any account, is making sure that you have the market covered, and you are bidding on all the right keywords.

Analysing the ‘Search Term Report’ within AdWords, gives you the chance to identify potential negatives, and is also one way to identify potential new keywords. Keyword insertion is good way of making sure that the most relevant ad is being shown at the right times, for the right searches, but are you sure that you have all the keywords you need?

For example many searchers these days make miss spells in Google, where potential customers made a typo in the search query. If your competitor has those typos as keywords, and you don’t, then you are at risk to loosing traffic and potentially conversions.

Thankfully, instead of manually adding these typos in to your account, Google are developing a new way to target these ‘miss spells’ and variations. They claim that soon, advertisers will be able to target misspellings, as well as “close variants include singular and plural forms, acronyms, stemmings (such as floor and flooring), abbreviations and accents” with their improved exact and phrase matching techniques.

An example featured in the Google Help centre include:

These changes are just some of the developments that Google is making to ensure they are at the forefront of search. With the competition building from the recent alliance of Yahoo and Bing and talks of a Facebook search engine, Google must continue to innovate when it comes to search functionality. It will be interesting to see the impact this new match type change has to CTR’s and conversion volumes.

Google Display Advertising Workshop

I was invited to attend the Home & Garden Display Workshop on 28th March at Googles new offices on St Giles High Street.

The colourful building was designed by Italian architect Renzo Piano and the new office interior is fantastic. Each room encompassed a new environment with different textures, seating, carpet and colours which created a very modern yet eclectic design.

The workshop consisted of attendees from both clients direct and agencies all focussed within the Home and Garden retail industries. The purpose of the event was to gather feedback from the participants on Google’s display network and build on this to inform us of the up and coming products and new functionality which could help us to enhance and improve the performance of our advertising on this particular platform.

Sales Attribution

The workshop kicked off with lunch which was nice and the opportunity to speak to some of the other attendees. Steve Fitton, Industry Manager –Media Solutions, then took us through the first presentation. He discussed the benefits of analysing the attribution of sales across all online activity via Google’s Multi-Channel Funnels. This is an important strategy we use for our clients to identify which platforms are influencing sales whether that be on a first, middle or last click. This type of analysis and the use of attribution modelling is becoming crucial to understand where and how to manage advertising budgets. Often brand keywords via search and direct traffic takes all the credit for the legwork which the non-brand terms and display advertising has actually created. This is just one of the tools that can be used to improve performance on the display network.

Contextual Targeting

He also discussed other areas of display including remarketing and similar users which everyone seemed to be seeing positive results from but the main subject to create some hostility was the contextual targeting. Many advertisers were finding their advertising appearing on sites which bared no relevance to which they had tried blocking and opting out of. The manual process to maintain this was taking its toll and many had thrown the towel in as they just were not seeing the results. It was explained that the display network should be reviewed separately to search but even when they weren’t compared the volume of sales were unable to justify the work or budget being invested. It was disappointing that Google didn’t really address this issue or go into any depth on best practice management or areas we could test to improve performance and therefore the issues raised were not really put to

bed in any way.

The Growth of Mobile

The second presentation (which should have been the first but was delayed) was from Peter Fitzgerald who gave us some obscure analogies, however presented very well and had everyone’s attention. He explained how important mobile was becoming and using an example of Kiddicare’s mobile landing pages demonstrated why separate mobile strategies must be targeted to mobile devices in order to generate ROI. He discussed some statistics on digital trends within the retail sector and then put a Q&A to the group to discuss.

Peter Giles, YouTube and Display Sale Manager then took the stand after a brief coffee break. He looked at some case studies and examples of how the display network has worked successfully for clients such as Groupon and ShoeDazzel, however for a UK workshop it would have been nice to see some UK company case studies!

One to One

The final session was by far the most useful and successful, one on ones with our Google contact, ours being Rick Jones, Industry Head for Retail. This allowed the opportunity for feedback on the session as well as a deep dive into our account to address any areas where we can improve, enhance or add some display strategies.

Overall the workshop had promise but I still feel I left wanting more detail and more insights over and above what I pretty much knew already. Hopefully the feedback given allows Google to take the next session to the next level.

As website owners, our main collective concern was the loss in analytics data that would result – specifically that the keywords used to find our websites by logged in users would be grouped under (not provided), but we were told by Google that on average, only 10% of organic visits would be affected.

Initially the secure search feature rolled out in the US and the impact of most UK based websites was minimal, however it soon arrived in the UK and I was curious to see what impact it’s had now things have settled down.

First let’s look at our own website, digital-clarity.com

Taking only organic visits, we’ve split them out into branded (where our name or variations are used), non-branded and “(not provided)”.

What we find is that rather than the 10% predicted by Google, it is 25.61% – just over a quarter of our organic keyword data is hidden.

Given the nature of our services and our audience, we expected to see slightly higher than 10% as many of our visitors are likely to have and use Google accounts, but 25% was well in excess of this.

I decided to take a look at some of our client’s websites to see how we compare and what sort of percentages you might see in different industries.

E-commerce

When looking across e-Commerce websites we found the following averages for branded, non-branded and not provided.

The domination of non-branded traffic is not too surprising – if the website is indexed well and makes use of relevant keywords, you would expect the majority to come through on product names or variations of them. For example, Tesco may hypothetically receive organic traffic to the Heinz Beans product page using keywords such as “Buy Heinz Beans”, “Baked Beans”, “Can of Beans”, “Heinz Beans Can”, ”Small Heinz Beans” etc. while the number of ‘Tesco’ variations you can use are limited.

“Not provided” averages out at 5.64% which is well within the 10% originally estimated by Google.

Online Software / Services

The following is an average of websites that offer either online software or services that can only be accessed online.

We found a large percentage of visitors from organic search to be using brand terms, which you would expect when the majority of visitors are returning to use the service.

Again the average percentage of not provided traffic is well within the original Google estimates at 5.53%.

Automotive

Websites in this category are a combination of manufacturers, retailers and service suppliers which make the figure given for non-branded terms a little unreliable – some websites are large and conducive to ranking for large volumes of keywords while others are more focused, however for the purposes of this post we’ll focus on the “not provided” traffic.

While higher than the previous industries, the figure of 6.42% still falls within the 10% or less estimate Google originally gave.

So What Does This Mean?

From the relatively small data sample available it would appear that most websites retain the vast majority of their keyword data for analysis, but for websites who’s visitors  are technical, in marketing or internet ‘power users’ there’s a good chance a larger percentage of them will be visiting while logged into a Google Account and therefore coming through Google secure search.

Recent studies suggest “not provided” accounts for 18%+ across many tech websites, while healthcare can see 13%+ as of January 2012.

Will We Lose More Data?

Given that Google is on a constant drive to attract new users it is fair to imagine the percentage will indeed grow as time goes by and more people remain logged into a Google account whether they’re a user of Gmail, Google+, Docs or any other service.

It’s also important to keep in mind that Google could at any point make secure search universal, meaning you will no longer need to be logged in to be taken to the secure version. This wouldn’t be unprecedented, the same process of US logged in > World logged in > universal rollout has happened with personalised search and other enhancements in the past.

Additionally there are many reasons why giving keyword data away is not in Google’s interest, some of which are:

• It can be used to manipulate search results through aggressive and questionable SEO techniques
• It’s a business that relies heavily on PPC as a revenue stream and like any business its overriding mission is to increase revenue for shareholders – giving away keyword data for free allows website owners to survive without advertising.

Firefox & Google Secure Search

Last week Mozilla announced that, assuming no flaws or bugs were found, Firefox would use Google secure search by default in its built-in search box features, regardless of whether people are logged into Google.

As of February 2012, Firefox was the most popular browser on the web (although losing ground to Chrome) with a 36.6% market share – for those of you keeping up that’s potentially 36.6% of your keyword data being lost to “not provided” once the latest versions are released and downloaded, which makes for worrying times for website owners that rely on keyword data to analyse performance in organic search.

Chrome & Google Secure Search

At present Chrome does not default to secure search for non-logged in users but it would not be surprising if this change was made, especially when a rival browser has done so already. However this may be delayed until a possible universal rollout for all Google searches.

In Summary

Google has little to lose if website owners are unable to see keyword data for its users and plenty to gain, so it would be a little naive to not at least entertain the possibility that a complete blanket blockage for all non-paying customers could come into effect in the next 12-18 months.

I can personally envisage a situation where such data is only available to either PPC customers or customers of a paid version of analytics, or select technology partners where Google has a stake.

However for now we’ll continue to work with the data we have with one eye on the future.

On May 26th 2011 a new EU originated law came into effect that requires website owners to make significant changes to their sites and may fundamentally change the whole web browsing and shopping experience for everybody.

This Cookie Law is an amended privacy legislation that requires websites to obtain informed consent from visitors before they can store or retrieve any information on a computer or any other web connected device.

As of May 26th 2012 this law will be in full effect in the UK.

Tom Collinson & Reggie James have put togeather an extensive post that explains what the law means, the implications and what steps you can take to ensure compliance for your website.

The EU cookie law: What you need to know

You may have noticed a lot of news recently concerning how personal data is used by large internet publishers such as Google, Apple and Facebook, for either improving service or as a pervasive and targeted tool for advertising.

Recently, Digital Clarity sent a newsletter highlighting the changes with Google’s Privacy Policy. It seems that the desire by the internet giant to simplify and blend some current 60 privacy policies into one is (currently) falling fowl of EU regulators…

View EU Cookie Law Guide on Startups.co.uk

If you have any questions regarding the EU Cookie law, compliance or assistance in carrying out a Cookie audit of your website please get in touch on +44 (0)845 388 4071 or email info@digital-clarity.com.

How is this going to benefit business that advertise on PPC?

This partnership will bring a number of benefits for search marketers, giving better access, better control and hopefully, better results for PPC campaigns.
Some benefits include:

• A single platform to reach a much larger audience than either separate platform could provide
• Improved compatibility with Chrome, FireFox and Safari for adCenter
• More control over the targeting of devices being used to search (mobiles, PCs and tablets)
• Yahoo! advertisers will now be able to reduce wastage with negative keywords
• Advertisers can optimise their marketing efforts by accessing social media data.

But it’s not just that simple. Whilst this merger will result in a larger audience, it also results in an inevitable increase in competition. This means you will have to be prepared to adjust your strategy to accommodate it.

Is this going to affect PPC advertising budget?

Yes and no.

Once the transition has been completed, both sources of traffic will be accessible from a single advertising platform.
If you are currently advertising on both platforms it may be as simple as combing the daily budget of both. However if you have only been advertising with Bing, increased investment may be needed to capitalise on the merging traffic.

Is this going to affect your bidding strategy?

Most likely.

At the moment the cost-per-click (CPC) of keywords can differ from Yahoo! and Bing, but once the merger is complete this will change to give a single cost per keyword. With this in mind, it’s important to remember that an increase in competition could lead to an increase CPC prices.

Report data is customisable and can be scheduled to run automatically if you regularly need to run the same report; however the process of running report data is different for AdWords than it is for adCenter. The following is a quick guide on how to run reports from each;

Google AdWords

1. Choose Your Campaigns

Once logged into your AdWords account, to run a campaign report you need to ensure you are in the campaigns tab both in the top menu and the sub menu. You have the option to create filters to view certain groups of campaigns; this is found in the ‘filter’ button just below the sub menu. Alternatively, you can run an individual campaign report by selecting the campaign you want to view from the list on the left-hand side.*

2. Choose Your Metrics

To choose which metrics you would like to be included in your report, click the ‘columns’ button (found next to the afore-mentioned ‘filter’ button), and simply add and/or remove metrics from the list and click save.

3. Set the Date Range

You need to ensure that you are looking at the time range you require, you choose this with the date ranges in the top right of the page.

4. Set Graph Metrics (optional)

You also have the option to change the metrics displayed in the graph by clicking on the graph symbol next to the columns button and choosing your desired metrics from the two drop down menus.

5. Set up Scheduling (optional) and Download Your Report

Finally, with all your settings in place, you can download the report by clicking on the download symbol, found next to the graph symbol. The menu which comes up allows you to choose your file name and format, and gives you the option to schedule the report to run automatically in the future. Once ready, click the ‘create’ button.

*You can run reports for ad groups and keywords in the same way by selecting the relevant tab in the sub menu. For all ad groups or keywords, ensure you have selected all campaigns in the list on the left hand-side, or you can run these reports based on filters or for an individual campaign as before. The dimensions tab allows you to run data by time segments, i.e. daily, weekly, monthly, yearly etc.

Microsoft adCenter

1. Reports Tab

For adCenter, reports are run from the reports tab in the top menu. Once in the report centre click on the ‘create new report’ button which will direct you to the settings menu for your new report.

2. Choose Report Type and Date Range

Under ‘basic settings’, you can choose the type of report you would like to run, for example, a keyword performance delivery report or an age and gender targeting report. You choose the time segment and date range here too.

3. Choose Your Campaigns and Metrics

Under ‘advanced settings’, you can choose whether you want the report to cover all accounts or you can select certain accounts, campaigns or adgroups. You are also able to change the columns and layout of the data tables to display the metrics you are interested in by clicking ‘change columns and layout’ and adding or removing the metrics on the menu which drops down.

4. Set up Scheduling (optional)

The final ‘templates and scheduling’ section allows you to change the name and format of your report and choose whether you want to schedule the report to run automatically in future.

5. Create Your Report

Once you have finished the settings click the ‘create new report’ button.

Summary

Both adCenter and AdWords allow you to look at analytics data within your ppc account, allowing you to see the location your visitors are coming from for example, or to see not only how many conversions there have been, but also the type of conversion completed (e.g. a registration or an uploaded CV etc).

Reporting is certainly simpler to do in Microsoft adCenter with everything in one place (which as a beginner I certainly appreciate!), it’s just a shame there isn’t nearly as much data to look at compared to in AdWords, as Google dominate the search engine market.

I’m sharing this mainly to see if there are improvements that can be made so I would welcome your ideas, opinions and feedback.

Before we look at how the tool works let’s set it up.

The Spreadsheet

You can get the spreadsheet tool here. But you’ll need to create your own copy so you can change the settings and keep the data private.

http://bit.ly/rZzW2k

Within the document is a sheet named ‘Settings’ where we’ll need to input the login details of your Google Analytics account and the source and medium for the traffic data you’d like.

I’ve included a key to show examples of medium and source settings but by default it’s set to call organic visits from Google.

Add Profile ID Keywords

Firstly you’ll need to get the profile ID for the analytics profile you wish to get the data from. Once you’ve selected it from the profile view the ID number will appear in the address bar  you’ll want everything after the ‘p’.

Finally, replace Keyword 1, Keyword 2 etc. with your desired keywords.

The Formula

I’ve used the following formula to pull in the Google Analytics data from the account we’ve specified in the settings sheet. Hopefully by breaking it down you’ll be able to not only understand but improve it and get around some of problems the limitations I’ve found.

=getGaMetric($B$6,”visits”,B17,C17,CONCATENATE (“ga:medium==”,Settings!$B$5,”;ga:source==”,Settings!$B$6,”;ga:keyword==”,A10))

=getGaMetric()

Tells Google Docs what the formula is to do, in this case it will request metrics from a Google Analytics account.

$B$6

The cell that holds the analytics profile ID

“visits”

The metric we wish to fetch

B17,C17

The start and end date of the period we wish to call data for

CONCATENATE()

I was unable to write a formula that could pull in the keyword automatically and fetch the data without doing it in fragments. I would welcome your ideas on improving this.

“ga:medium==”,Settings!$B$5

Fetch visitor data where the medium by which they arrive exactly matches the value in cell B5 on the settings sheet.

“ga:source==”,Settings!B$6,”

Fetch visitor data where the medium set above is true and the source exactly matches the value in cell B6 on the settings sheet.

“ga:keyword==”,A10

Fetch all the data that matches the medium and source specified where the keyword matches exactly the value in cell A10.

Issues

Data Can Be Slightly Off

It’s logical to think that as the data is being pulled directly from Analytics the figure outputted should match exactly the figure given within the Analytics interface yet it seems to always be less, albeit by a small amount.

It’s possible that Google Docs works on a different time zone to that set within the analytics account but I’m unsure.

Too Many Requests

I’ve only included 4 keywords as 4×12 months = 50 requests for analytics data at any one time. This seems to be the limit as if you add anymore it returns an error.

Does anyone know if 50 is indeed the limit or if there’s a way to get around this?

Changing The Year

This method requires a start and end date be specified for the data you wish to collect, hence the table January to December. Is it possible to have the year held in a separate cell and pulled in to make the date valid?

I have tried this method and while it’s entirely possible to concatenate two cells to give a valid date format the getGaMetric function does not seem to like it.

Here’s the spreadsheet again – please get in touch if you find it useful or have any suggestions on improving the usability.