Last month I spoke at SAScon in the UK on the topic of getting actions from analytics. It's something that I feel quite strongly about so I wanted to write a post on here to talk about it a little and expand upon what I talked about at SAScon.

The key question I want to answer is -

"What key metrics should I be tracking?"

I can break the answer down pretty quickly by saying the following -

1. Metrics that are actionable - why measure stuff you can't do anything about?
2. Metrics that help you get stuff done - using data to fight internal battles is vital
3. Everything else - meh, wouldn't worry about it too much!

Whilst I love diving into data and getting into the how and why trends appear in analytics, I have to remember that my time is limited, in particular when looking at a clients analytics account. I need to understand as much as I can, but ultimately I need to focus on the metrics that make a difference. Also if you are a bit of a newbie in SEO or analytics, the amount of data available to you can be very overwhelming. Far too many people make the mistake of focusing on the wrong metrics which can be costly to a project. It can sometimes look like you are doing a bad job if you are measuring the wrong metrics.

Metrics that are actionable

If a metric isn't actionable, then its just fluff that fills up space in a report and doesn't mean very much. Whilst metrics can be interesting, there is a difference between a metric that is interesting and one that is actionable. Below I've listed a number of metrics that I define as actionable.

1. Beat Panda - Measure Content

Whilst the Google Panda updates of 2011 have looked to address a number of quality issues, one of the key ones in my opinion was the amount of good, quality content on a page. Many pages of thin or duplicate content across your site could potentially cause your entire site to fall foul of Panda. One way to assess this on your own site is to use custom variables to measure the content on your page. This works very well if you have an ecommerce site or any site where you collate reviews:

Action to take -

• Measure the number of reviews on your pages and see if pages with little or no content are as visible in Google

If you want to learn more about using Google Analytics custom variables, checkout this tutorial video.

2. % Visitors who view product pages

The reason this is important is quite simple, no view of a product page = no product sale.

This is one that is usually taken for granted by many ecommerce sites. They assume that because they have lots of product pages, that visitors will automatically find them. However there can be barriers to a visitor getting to your product pages, for example poorly designed navigation, poor internal site search, unclear category names etc. Also lets now forget that many ecommerce sites have other sections of their site such as a blog, a news section, buyers guides, videos and help pages, so a visitor can easily get distracted by one of these and end up leaving your site without ever seeing a product page.

Actions to take -

• Improve internal site search
• Improve navigation menus

Pro tip - make sure your internal site search works for both plural and singular product names. A classic mistake I've seen over the years is exact matching of queries to product names which means either singular or plural returns no results.

3. % of people who search your site then exit

If a visitor has made the effort to type in what they are searching for on your site and still left without buying, then you should take a closer look into why. So many internal site searches are not only badly designed, but also return bad results. If your website is like this, then you are losing sales. Actively improve your search results pages to make them more relevant and useful. it still amazes me that SEOs obsess over Google SERPs, yet neglect to take a look at the quality of their own despite having control over them!

Actions to take -

• Test it and make sure it actually works
• Do the results you see make sense
• Include special offers or discounts in results
• Enhance results with images - see Apple for a good example:

If you haven't got it setup on your site yet, check out this information on setting up site search for Google Analytics.

4. Page load speed

If it isn't already clear to you, Google is obsessed with speed. I can see why, a faster web is better for everyone and we all get frustrated if a page takes longer than a few seconds to load. Google have confirmed that site speed is part of the algorithm, albeit a small part. They've also made it measurable in Webmaster Tools and more recently they made it available in Google Analytics.

Now that we can measure load speed in analytics, it becomes much easier to see the results of the actions you take. Its also easier to see which pages of your site are slower than others, which allows you to try and work out why and fix the issues.

Something I wanted to point out here too is that you shouldn't be worrying about site speed because it forms a small part of the Google Algorithm. You should be worrying about it because it is important to your visitors, they will not come back to your website if they are left hanging around waiting for the page to load.

Actions to take -

• Look at which pages or sections of your site are the slowest
• Focus on improving the speed of your money pages
• Bug your developers and educate them on why its so important

If you want to learn more about what you can actually to do speed up your site, take a look at Craig Bradford's site speed for dummies part one and part two on the Distilled blog.

5. Average Order Value

This metric is all about squeezing as much revenue as you can from each order. Quite often, an SEO will start a project and the first thing they do is assume they make more revenue for the client by getting more traffic. Whilst this can be true, the quickest wins can often come from taking a good look at ways to get more value from the existing traffic.

There are actions you can take though if you are looking to drive traffic that will increase your order value.

Actions to take -

• Add up sells as much as you can to the buying process - Amazon are awesome at this
• Look at keyword trends to see which ones drive a higher order value
• Look at traffic sources to see which ones drive a higher order value

6. Measure SEO Variables

Similar to point one above, you can also use custom variables to track various SEO variables of your website such as -

• Pages tagged with Rel=canonical
• Pages that are linked to sitewide
• Tag certain "types" of content across your site

Action to take -

• Measure things such as the above and see how these pages perform in search results

7. Completion of Tasks

This one falls outside Google Analytics but I still wanted to measure it as it can be a great way of getting quick feedback about your site. I'd recommend taking a look at 4Q survey for this one, they offer a free trial so you can give it a test drive and see what you think. Avinash wrote a pretty in depth review of it on his blog if you want to read a bit more into it.

Quick sidenote, if you are into Analytics and not following Avinash on Twitter or reading his blog, you should be! He is also speaking at Mozcon in July and I've heard great things about his presentations.

Actions to take -

• Ask your visitors questions that matter using this software and feed the answers back into your strategy

8. Share of Search Landscape

I really like this metric. The main reason I like it is that it is a metric that anyone can understand, even people who don't get SEO will understand this. This is because its very similar to a traditional marketing metric of market share, so it can be used when demonstrating SEO growth and targets to your clients.

Here is an example of how this may look (dummy data!):

9. Form Field Tracking

This is such an easy one to action and can make a big difference to your conversions. If you are able to identify fields that are consistently confusing customers and perhaps making them abandon the form, then you are able to look into why and remove these barriers to conversion. You can even take a step back and ask yourself if that field is even required at all.

Action to take -

• Identify form fields that are causing problems and either remove them or improve them

Pro Tip - you can use Google Analytics events to track which form fields are causing problems. Here is a good thread over on the Analytics help forum which points you in the right direction for setting this up.

10. Branded vs Non-branded Search Traffic

If you are an SEO, or you employ an SEO agency, you should be measuring this metric. The great thing about measuring non-branded traffic is that pretty much 100% of the credit can be given to SEO efforts. Whilst credit for branded search increases can be credited to a number of marketing channels such as PR, offline advertising or online display advertising.

Action to take -

• Measure non-branded keywords and see which ones convert best, then focus on increasing rankings for these

11. Conversion Rate

I'm pretty sure you are all looking at this metric, but I bet that some of you are not using it properly. By not measuring it properly I mean that you are probably looking at this overall figure:

You can't take any actions from this figure. In order to take actions, you need to segment by relevant dimensions such as traffic source, type of visitor, location, keyword etc.

Action to take -

• Segment conversion rate by best performing dimensions and invest more resource into those dimensions

Metrics that help you get stuff done

The following metrics are focusing more towards helping you get stuff done. You should really make sure you use all available data to fight internal battles and give the business case to effect change at your clients company. Make sure that you backup your recommendations with solid data, however be careful not to confuse key stakeholders with complicated stats or fancy graphs. Just give them the data they need and visualise them in a meaningful way.

I'm going to give a couple of client examples here, whilst I can't give away the client themselves, hopefully these examples will demonstrate the power of data in getting things done.

1. Client didn't appreciate the value of content

A client of mine launched a microsite which had lots of good content on it about their upcoming releases and products. When their main site went live, they took down a lot of this content. This was despite our best efforts to stop them! We saw traffic drop slightly and growth wasn't as fast as it had been in previous months. After a few months (yes I know!) we finally got the content put back up and we pretty much instantly saw an increase in traffic.

So the next chance I got, I did a short presentation to one of the company directors who had the power to make things happen at the client. He wasn't involved much in day-to-day SEO but still had the power to influence the rest of the company's attitude towards SEO. I showed him a simple bar chart showing the traffic drop and the subsequent increase.

It worked a treat and he now understands the power of content and how it effects organic traffic. This means that future tasks that involve content won't be buried deep at the bottom of the developers list and will be slightly closer to the top. Influencing change ftw.

2. Developers didn't like SEO tasks

One client of mine has a team of developers who are probably the smartest I've ever worked with. However they'd previously worked with SEO companies who had got them doing work which didn't make any difference to traffic or revenue. These guys are super busy and didn't want to spend their valuable time on tasks which didn't make a difference. This made them a little resistant to future SEO tasks.

So when we had a situation where it was possible for Google to crawl literally billions of pages on their site, most of which we didn't really want Google spending their time crawling. After some work, we managed to make some changes with the help of the developers to solve this problem and make Google focus on the important pages and spend more time crawling those.

A month or so later, we saw some very good results both in terms of increased traffic and could also see through server logs that Google were doing exactly what we wanted them to do.

With the help of the client, these results were shown to the developers and they were given credit for the increased traffic. They were able to see the results of their work which makes it much easier now to get them to work on SEO tasks.