Author Archives: alicereidmorgan

Keywords for natural search – what’s the optimal number?

Posted on by
Reply

Quite a few people have been asking me this question lately.

Experience I’ve had over the last few years of running a centrally-managed pan-European SEO programme for Chevrolet Europe has taught me a lot.

We need to focus our energies

With so many markets on the programme (19 in 16 languages) and such a small team, we can’t afford to waste time or effort. We need to focus on what will bring us results. I can confess that I’ve learnt this the hard way…

SEO keyword lists are not the same as paid search lists

Natural search keyword lists should only ever represent an indicative target. One look at Omniture, or whichever analytics tool you’re using, will show you that there is a multitude of keywords people use to find you. So we now no longer include every possible permutation of every brand term in our lists.

This kind of coverage is actually important – to start with at least – for paid search. But natural search requires a different attitude when it comes to keyword lists.

So, what do I mean by keywords that will bring us results?

Our SEO agency Performics use various tools to identify keywords that – should we rank for them – would give us a traffic increase. This applies to both brand and unbranded keyword selection. It’s a strong and pragmatic approach.

Impact on content

If we’re not ranking for a term that the agency have identified as a potential traffic booster, we look at developing appropriate, relevant and useful content around that term, so we can rank for it. And, if necessary we build links too.

A good example of this the work we’ve done at Chevrolet with our unbranded category car pages. Here’s an example: http://www.chevrolet.es/nuestros-modelos/coches-de-7-plazas//. We’re also looking at other content development at the moment.

So, what of the results? Pretty damn good so far

I’m going to focus on the example I gave in the previous paragraph. By ranking no.1 for “7-seat car” (and related terms), we attracted an incremental 34,000 visits to the 7-seat car page on the Chevrolet Spain website in Q4 2011. I’ll say that again: incremental traffic. For a brand like Chevrolet Europe, with very low media budgets, this is an important step in generating awareness and getting our cars in front of people who would otherwise not have considered us.

I will admit to cherry-picking the best result we have. Not all our markets have performed like Spain, but in every market we’ve adopted this policy, we have seen a steady increase in traffic that goes back to a small, targeted set of keywords.

Paid and natural search alignment

And it’s part of our paid/natural search alignment project. All the unbranded keywords we’re targeting are expensive (but effective) in paid search. If you’ve read my previous posts, you’ll know that we’re testing an active bid management system which will allow us to determine the optimum relationship between paid and natural search. Spain is one of our test markets and the results are looking extremely positive.

So far, it’s showing that we can radically improve traffic and conversions to the site by being present in both paid and natural search. Performics will be writing a white paper on this test later in the year.

Why is bad to target a large amount of keywords for natural search?

OK, back to the beginning. This is the question that people often ask me too, in relation to the main question “what’s the optimal number”?

Well, for starters, it’s very wasteful. But much worse than that, is that it can be very content-destructive. If you’re under pressure to rank for a huge number of keywords, the temptation is to produce content that’s not consumer-focused (i.e. it’s not useful for the consumer, isn’t easily readable and so on).

So, my advice is…

…keep the lists to under 150, including unbranded terms, be highly selective and write content that satisfies consumers’ needs first. If you get it right for consumers, you’ll generally also get it right for search engines.

There is also a social element. Where you think your content is sharable, or likeable, do give people this opportunity. But use it judiciously. Nothing worse than seeing “Likes 0″ on a page.

Let me know what you think about this article!

SearchCenter+ test is underway

Posted on by
Reply

After a couple of frustrating weeks, our test is live in the UK, France and Spain. Credit to Sam at Performics, Rina at Chevrolet, Frank at Ignition and the team at Omniture for overcoming the last few difficulties to get started.

We’re testing our hypothesis (that you can dial back spend on paid search when you’re ranking well for natural search) using SearchCenter+. It’s one (big) strand of the integration we’ve driven between paid and natural search at Chevrolet Europe.

More on this as we get results!

Pan-European Linkbuilding

Posted on by
Reply

In my last post I outlined plans for Chevrolet which are about to get underway for testing an active bid management tool – SearchCenter+. I mentioned in passing that we’re doing linkbuilding to boost our natural search rankings for generic keywords, allowing us to get into the position of being able to adjust our paid search bids. I’ll keep you posted on this. (Inevitably, there is a hold up – not of our making.)

Frustrations aside, I want to give a bit more of an insight into what we (that’s Perfomics, Chevrolet’s SEO agency in the main – and myself) are doing on the linkbuilding side, for those of you interested in this topic.

Chevrolet is a real branding oddity. Amazing spontaneous awareness for the Chevrolet brand on the one hand, with virtually zero awareness for the cars we sell in Europe on the other. Back in 2008, our digital agency, Ignition, ran some consumer testing. A mix of Chevrolet and non-Chevrolet owners was recruited. When we asked the owners if they could name a Chevrolet model other than the one they owned, the answer was a prettty unanimous no, they couldn’t. This has changed somewhat and awareness is steadily growing, but from a very low base. Currently Chevrolet’s market share in Europe is just too low to get serious “on the road presence” (though, judging from the cars I see around London, I believe this may be changing too). And then there’s the question of money. There isn’t enough for us to break through into broadcast or any other meaningful interruptive media. We can’t book the Champions League half-time break. Especially not right now.

That’s all a rather long-winded way of saying that we need to find other, more cost-effective ways of gaining awareness among people who don’t already know us. And one of these is to do linkbuilding.

A Starcom stat I picked up about a year ago claims that 88% of the 7 billion automotive searches each year are generic. That’s quite an opportunity. Another eye-opening view comes from Searchmetrics, who have devised an algorithm which estimates the traffic increase if your website ranks no. 1 on Google – by keyword that does so. The bottom line is that you don’t have to rank high for many popular generic keywords to boost traffic considerably.

But why linkbuilding? What’s the matter with regular SEO techniques? SEO-ers among you will already be well aware of the answer. Ranking for generic keywords using content alone is virtually impossible, especially as we want to target quite a range of them. It’s obvious that generic keywords are much more competitive, so it’s always been quite difficult, but Google has made it even more so with its last few algorithm changes. Mayday in 2010 was the real deal-breaker and this year’s Panda has just consolidated it. Google’s focus is on the delivery of excellent content for consumers: not helping big brands rank well in natural search. What is required are external links which confirm the relevance of our content. And Performics have a team of people patiently, persistently nurturing bloggers to do just this for us in 13 markets across Europe, including smaller ones like the Czech Republic and Austria. We don’t get the instant gratification of paid search. It takes a considerable amount of time and effort, but once in place, it’s sustainable visibility.

Finally, I just want to return to one of the topics in my last post about our one-search experiment. Our linkbuilding strategy has always been the first step along the road of establishing a symbiotic relationship between paid and natural search. And as such, our linkbuilding keyword recruitement criteria keywords are paid search keywords we’re already bidding on, which we know are effective at driving traffic and conversions, but also very expensive. And that, along with the Starcom stat, is our big opportunity.

Paid and natural search integration – testing SearchCenter+

Posted on by
2

Everyone working in Search should want this: make paid and natural search work together in a symbiotic way that leads to spending either less money on paid search, or at least spending it more efficiently.

It’s something we’ve been trying to get going for some time at Chevrolet Europe. It all sounds so simple: reduce spend where we consistently rank well on natural search and redeploy the budget. Let me tell you: it’s not simple, especially when you have 20 markets to consider.

Convincing new technology
So I didn’t have that much hope when I went along to a seminar hosted by Adobe and SearchMetrics in London in September. But I came away from it convinced that the technology we’d been shown could work for us. SearchCenter+ pulls in SEO analytics from SearchMetrics to create a platform that allows paid search positions to be automatically adjusted according to the natural search position and other competitor factors. More on this active bid managment tool.

A search industry first for Chevrolet
Tamara Warren, Strategic Digital Marketing Manager, and Rina Patel, Website Marketing Manager (who knows a hell of a lot about paid search) were very enthusiastic about the idea too. Such was the momentum we created that the test will start without delay in January 2012. Managed by our SEO agency Performics, it will be the first of its kind in Europe because – right from the start – we’re testing in a mulitple market environment: Spain, France and the UK.

Our “one search” keyword selection methodology put simply
For a few months now we’ve been running active linkbuilding campaigns across 13 European markets, targeting mid-funnel generic keywords. The selection criteria for these keywords were paid search keywords with:

- high search volume (opportunity)
- but high cost per click and cost per conversion

This approach lies at the heart of our “one search” test because right from the start, our keyword selection has been tied to our paid search activity. It’s the natural search positions of these linkbuilding keywords which we hope will provide the right testing conditions.

Updates as we go
I will publish updates as soon as we get an indication of whether the theory of active bid management can be proved or disproved.

An Alice Morgan blog? Well, it’s about time…

Posted on by
1

I’m always full of admiration for other digital marketing professionals for their ability to work and blog about their work. I’ve benefitted hugely from reading other peoples’ blogs – and at the same time always thought I could also add to the debate and contribute. After all, I am involved in some really interesting projects that other people might want to know about. But then there’s the dog, the kids, the extra maths, the tennis coaching the shopping. Somehow – I can’t imagine why – I never have found the time.

Well, I’m making it now and I hope you will find this blog interesting and I hope – soon – to be joined by another key partner, who will also contribute her experiences alongside mine.

In this issue I would like to tell you about the interesting “one search” project that I’m working on for Chevrolet Europe.

If you want to find out more about me, there’s a short summary in the “About” section or feel free to check me out here: Alice Morgan on LinkedIn