Search marketing: How to best benchmark and measure ROI

While some marketing disciplines such as social media marketing are notoriously difficult to measure performance and ROI, search remains one of the more quantifiable practices.

In its Marketing Budgets 2013 survey of 800 companies last year, Econsultancy found that few organisations were unhappy with their ability to measure ROI from search.

89% rated their ability to measure paid search ROI as either good or OK, with only 11% deeming it to be poor; while 83% of those questioned rated their ability to measure natural search (SEO) as good or OK, with only 17% ranking it poor.

Nonetheless, these findings reflect that the situation regarding the measurability of SEO ROI is markedly different from that of paid search.

In his contribution to the SEO Now ebook, Marcus Tandler, partner at Tandler.Doerje.Partner, notes: “It’s not like AdWords, which has a direct and attributable ROI. With paid media, marketing managers have gotten used to the mindset of spending $100 and looking to make $120. If it works, then they scale up the budget and there is always a comfortable reporting metric on their investment.

“With SEO the situation is that you do the best you can, as informed by the available knowledge and data, and wait to see the results. There are a lot of variables which you cannot influence, and the results come in much more slowly. Google needs a certain amount of time to crawl pages, and it can be very difficult to manage expectations for an SEO campaign.”

He continues: “It’s made more complicated by the fact that an SEO win is not necessarily pushing a site to the first page of Google, but in optimising the entire search experience, taking user intent and the delivery of the best results into consideration when operating in the organic space.”

Danny Sullivan, founding editor of Marketing Land and Search Engine Land, adds: “It’s easier on the paid search side because you can track down to the key word level and have that passed along to you so that you know someone came to you and you can target that and keep track of them.

“On the SEO side it’s also now harder because Google has been stripping out the terms for over two years, so when someone arrives at your site you don’t really know what terms they used to reach the website so it’s hard to attribute them. You can get that in other ways but you won’t necessarily know the individual person that did it.”

A picture of performance

But with over three-quarters of businesses reporting that they’re satisfied with the measurability of their SEO ROI, the majority of organisations are clearly content that they can get an accurate enough picture of their performance. 

“You can still tap into tools like Google Webmaster to get a sense of how your site’s performing overall. And analytics tools will tell you what’s going on with search in general. You might not be able to do as many things in terms of trying to track the lifetime value of a particular person, but there are still things you can do to get an idea of what’s going on with your SEO in general.”

Elaborating on this point, Alexandra Tachalova, brand manager at SEMrush, recommends that companies should be monitoring conversions and revenue, quantity and quality of backlinks and rises in brand mentions as a matter of course when it comes to SEO.

“It’s really important to track the positive changes in website rankings and also organic traffic,” she explains. “Another factor which should be measured in SEO is the conversion of traffic to sales, so traffic growth isn’t the only real result of a successful SEO campaign. Finally, a set of reports should be done to see the increase and quality of backlinks and the rise in brand mentions.”

But Kelvin Newman, managing director of BrightonSEO and a tutor on the IDM Award in Search Marketing, believes that an even more granular level of measurability is possible.

“Most websites now have Google Analytics or an equivalent measurement system in place, and those tools are effective at separating out the sources of traffic that you’ve got,” he explains. “And if you’ve got your website set up in a suitably friendly way for the analytics, it can tell you exactly how many leads you’re getting or how much you’re selling. If you know how much you’re spending to get that traffic, you can understand what your ROI there is.”

Paid search measurability

When it comes to paid search ROI, things appear to be much more straightforward, with metrics such as clickthrough rate, conversion rate and cost per conversion all providing good indicators about your performance.

An obvious approach to measuring ROI is therefore clearly defining the objective that you are trying to achieve (i.e. enrolling to a webinar, lead collection, downloading marketing collateral), generating a conversion rate (telling you how many people completed this objective) and then calculating the cost per conversion ratio by dividing the number of successful conversions by total cost. The likes of Google AdWords and Google Analytics collect relevant user information that can be reviewed so that you can also compare the difference in the quantity and quality of conversions between traffic paid for by you PPC campaign and the ‘free’ traffic that you have generated by your organic search efforts.

A cost per action (CPA) pricing model can be another alternative for organisations using paid search, which is a model that allows the advertiser to pay for an action a user takes (lead generated, sale, download), rather  than merely a click on the advertisement.

But even despite all of this, things aren’t without their complications when it comes to paid search.

“The difficulty is there are always questions of incrementality and cannibalisation,” explains Newman. “So, if it increases was it the market that increased rather than necessarily better performance? And is your search doing better at the expense of other channels – maybe you’re running a paid search campaign, would the people that are converting have done so anyway, just perhaps via another channel were you not running that campaign.”

Furthermore, AdWords continues to become more complex, which also makes it harder.

“The system has become bloated, opaque and hard for advertisers to use effectively,” says Rhys Williams, managing partner at digital consultancy agenda21. “The number of updates is unrelenting and requires constant attention… The majority of these increase the volume of clicks and therefore revenue to their business.  Now, I have no problem with that and I guess neither do the shareholders.  However, it has made the system so complicated that unless you are an expert user, it’s very hard to get the best from it.” 

Benchmarking

For this reason, Williams suggests that marketers should focus on benchmarking current performance and connecting PPC investment directly to their strategic business objectives. Supporting this, the Institute of Practitioners in Advertising and the Internet Advertising Bureau have both taken an active role in supporting an independent benchmarking process over the last two years and have been helping to establish metrics, KPIs and best practise.

Newman agrees that this kind of benchmarking would be sensible.

“Never has it been easier to understand how your business is performing, yet in many cases people are still very keen to get more people into the top of the funnel and perhaps are not necessarily spending as much money in terms of getting people through it,” he notes. “Say a business gets 1,000 visits to their site which turns into 100 sales, their instinct is to say, let’s get 2,000, or 200,000. But actually, in many cases, just by understanding the data they already have access to, they might be able to get to that 200 without even necessarily increasing the number of people coming to the site. This is just by understanding what areas of the site are failing, what could be tweaked slightly, where are people dropping out and leaving – all the types of questions which a free piece of software like Google analytics gives you access to.”

So what do businesses need to know about benchmarking?

“First off, make sure you’re collecting the right data – it sounds stupid, and it sounds basic, but you have to be collecting the right data and you need to be collecting it at a granular enough level," recommends Williams. "You need to be able to suck up keyword level data, so you need to be able to track from a keyword right the way through the process on your site, and what you should then be able to do is create segments of users, people who have bought from you, people who dropped out during the process, people that you can cross sell a product to. So you begin to create these segments, which you can then plug back into your search activity to start retargeting them. There is real value in having that kind of set up. A lot of brands tend to collect the data, look at it, but then nothing really seems to happen with it.”

Newman has a few final words of warning when it comes to getting benchmarking right. “There are studies and reports out there that can give you ideas of average bounce rates or average time spent on site, but that will vary from industry to industry, and even within the industry from website to website,” he says. “To use travel as an analogy, on many websites, what they want you to do is get booked as quickly as possible. So maybe on a flight search engine, having someone spend a lot of time on the site could be considered a bad thing and a failing of your site as they’re not navigating through and making that purchase as quickly as possible. Whereas on other websites which are perhaps selling more of a considered sale, actually it’s a good indication, as you’re expecting people to make multiple visits, browse round and explore your site. So it’s tricky to get a true benchmark there.

“Therefore, the best benchmark you’ve got is against yourself - you can look at how you did last year, how you did last month and benchmark against yourself.”