How “Stupid” is Behavioural Targeting?

With so much talk about Behavioural Targeting’s opportunities to us Internet Marketers, it’s worth a little reflection on the purpose of Behavioural Targeting and a few of the issues which need ironing out.



The Purpose of Behavioural Targeting

It really is quite simple; track a user’s interaction with your website, and serve ads, promotions or recommendations based on what others do.

These ads, promos or recs could be within the same site the user is on, on another website you own or any other website which accepts and subscribes to one of many behavioural targeting ad servers.

So the purpose is, with my customer advocate hat on, to enrich and personalise the user experience by offering the right message to the right customer.

However my marketer’s hat suggests it is a great way of increasing click through rates and conversion per user. It’s like a store assistant following a customer around holding the items they just looked at in the hope they might actually add the item(s) to their basket.

One could say this is quite convenient for a customer and perhaps a little persuasive (pervasive?). Maybe I will buy it after all since you hounded me so politely…

It could just as easily become quite annoying. Picture yourself window shoppping. You know you can’t afford this particular item but you’re just taking a look… like a holiday for example… 14 nights in the caribbean perhaps? So you leave the site and jump back onto your surf board across the internet…

Targeting is not child's play...

Targeting is not child's play... (image:aboyandhisbike|flickr)

Next time you check your email, you notice the banner ad displaying the exact holiday you were looking at. You ignore it but when browsing the latest news see the ad in numerous places… now you’re getting upset. Quite annoyed in fact. You feel teased.

Behavioural targeting is not exact science

Behavioural targeting for e-commerce that doesn’t use anything more than your clickpath data through a website is severely lacking in one crucial ingredient – what you actually bought. Whilst the last 5 people that clicked around a website like you did may have all purchased item X, you bought item Y. You’re different, unique in fact.

Extending intelligence

Even if the targeting software held data on your past purchases, how could it predict your next purchase? If a customer bought a variety of apparel such as jeans, shirt and shoes, how can you predict their next purchase? Aggregation and segmentation. Whilst the propensity to buy something is not an exact science one can identify triggers and signals in data to determine cluster behaviour.

I can safely predict what time I’m going to want to eat but I won’t necessarily know what I am going to eat. This magnifies the problem online – not only are we trying to close a sale, but best predict what they might buy in order to increase the basket value.

Combining a user’s click data together with history of purchase can only take us as far as predicting their behaviour based on groups of similar consumers.

Behavioural targeting vs personalisation

When crowd behaviour creates pigeon holes, not everyone’s going to appreciate it – even personalising content can become a broadbrush without various other magical ingredients…

So taking our annoyed subject, she won’t be enjoying a hotel break anytime soon.

So how could this form of behavioural targeting be improved?

The Conversion Funnel

Don't stalk your consumer!

Don't stalk your consumer! (image:tomconger|flickr)

This is an assumption, that our subject had indeed visited the website and would really have loved to part with her money, had she had some. However how closely do site owners think about the conversion funnel? This is where behavioural targeting can become really clever. Not only base a recommendation or banner ad based on behaviour AND past purchase behaviour but also on what step of the journey the customer may be on. If you’re targeting a user that has clicked around a particular hotel more than others and maybe even stayed at your hotel chain, have they stayed at this one? Is your brand functional or aspirational? Should you perhaps woo your prospect or are they simply after the best deal?

There is a deep level of intelligence we can delve into, to really uncover the depths within the consumer’s thoughts, their needs and desires, to fully develop a truely personal experience that betters the high street and then, right when we think we’ve mastered it all…. a law is potentially being delivered banning the exact things we believe will help bridge the gap between the high street and online experiences (to an extent)

Is behavioural targeting really just plain stupid? I think not. However businesses need to reassess WHY they’re doing all of this and listen to the consumer – as with any change there will be those that protest but if the consumer feels they’re getting more benefit than not, surely they’ll stick it out…


Think Vis 2010 – An Internet Marketing Conference To Attend

About Think Vis


The Think Visibility Conference

Think Visibility is a conference on points you might miss elsewhere at other online marketing events. Take advantage of well-known speakers and emerging talents and networking opportunities with industry experts worth listening to.

The Think Vis events are usually held on a Saturday to ensure you get the opportunity to attend – no excuses accepted!

Find out more about the Think Vis September Event or book for the next Think Visibility >

Think Visibility Golden Nuggets – September 2010

Take a look at some of these articles on Think Vis for a real taster:

Summaries of ThinkVis

Affiliate Marketing @ ThinkVis

Design, Usability & Conversion @ ThinkVis

SEO @ ThinkVis

Did you attend Think Visibility 2010? Any other interesting blog articles worth adding above? What did you think of it and how would you pitch this to fellow Internet Marketers?

SEO and Multivariate Testing – Be Aware

Multivariate Testing (MVT henceforth), the art of testing more than one page element in a given space of time simultaneously (almost) has proven its worth to me over the years on numerous projects.

Gone are the days of testing single page variants and instead we can now test numerous permutations of pages with various page elements in a fraction of the time required for simple A/B/n testing. However this does introduce an interesting element of concern to us SEOs, concerned not only with ensuring page content converts but also that the page performs at capacity as far as Search Engine Robots are concerned.

So with your MVT tool serving many variants of various parts of your website, have you considered the impact on your SERPs? Many of the MVT tools do recognise a bot from a web browser, but have you actually checked your tool? The effect that this could have is profound.

Take this example; your site is highly indexed and you’re running MVT iterations on your homepage. Your MVT tool is not segmenting robots out of the tests and the Search Bots see different layouts every day (because your site is visited every day). All that hard work you, the SEO guru, spent on optimising the homepage for maximum SEO performance may well be compromised and undermined by way of changing layouts and content.

Even if the Search Bots are being filtered out, what if the winning permutation of the [Home/Product/Insert Name] page reduces the effectiveness of your page in the eyes of the Bots? This is a crucial aspect of MVT that is often overlooked.

My advice? Unless you’re also responsible for MVT, ensure you are fully aware of what is being tested and the impact any new layouts will have on your SEO strategy. After all, increasing page conversion yet decreasing traffic may leave you spinning your wheels in the sand…

What are your experiences? Would love to hear of other real-life examples

Affiliate marketing – the long-tail approach

My biggest earning affiliate marketing website exists within a highly competitive sector: travel. The website itself targets the uk market which is a small playing field with a large number of players. Ranging from OTAs (Online Travel Agents), holiday providers, private and franchised accommodation providers, from hotels through to caravans, cottages, cruise ships and B&Bs and of course a mass of well run, fully established affiliate sites it could easily put many new players off joining the ranks.

My approach as with all projects is to start small, test the waters and then take a call on whether to continue or scrap it. I was close to scrapping this particular venture 4 months in but persevered to see average page impressions each month hitting 50k. With an average of 10k UVs (unique visitors) a month that is an average of 5 pages viewed per visitor per visit.

Traffic is acquired primarily through SEO (search engine optimisation). I have PPC (pay per click) campaigns set up ready to drive more volume but the website is still in learning mode, to find out which keywords are driving traffic and converting into a sale (for free).

The point of interest raised from analysing the keywords, is the level of brand traffic this website is attracting. That’s not to say it is outperforming the brands for which I’m promoting holidays for, but that I am plugging a gap in the SERPs (Search engine results page) for those brands.

This is a crucial area for affiliate marketers without a strong niche – it can be done. What the big brands don’t have time for is chasing that 20% of long-tail traffic when the 80% of short and part long-tail they already acquire is turning over the majority of their revenue… and this is where affiliate marketing can work with the brands to complete the ultimate goal of customer satisfaction.

The majority of my efforts have gone into on-site SEO to begin with. By creating a strong search architecture, it allows for the website to continually take advantage of a well structured, robot friendly infrastructure together with the ability to maintain the daily content release process without the need to concern oneself with the technicalities of SEO. It is for this reason the daily content management of the site is out of my hands as is the on-site SEO, leaving time for more off-site SEO. This is the key to enhanced and sustainable search engine positioning.

This has also given me a rich dataset of top performing keywords to build a successful PPC campaign from whilst earning its own marketing budget. Therefore in year 2 of its operation I aim to double income by both increased SEO traffic (domain age, authority etc) and highly targetted PPC campaigns…. and whilst that’s all going on I know a thing or 2 about conversion and retention I’ve yet to utilise.

I shall keep you ‘post’ed!


Online Branding & eMarketing Techniques

I’ve spoken to a number of leading companies in the ecommerce arena recently including retail, advertising, investment and travel. Most of the brands, which will go unnamed for reasons of trust and confidentiality, have a strong traditional ‘offline’ brand, well known as being leaders in their industry. However, they share 1 commonality – using their brand strength to drive online revenue, without too much focus on recreating the core brand messaging online.

What’s wrong with this if your company has continued to grow from strength to strength since early 2000? The problem is both of market saturation and losing out to your competitors. Using the stronghold that your brand has on a particular sector to drive online sales, will eventually mean you are losing out on a chunk of online sales of customers that do not know your brand well enough.

A true multi-channel sales approach considers online as more than an end-point and instead uses emarketing as an integral part of your marketing strategy. Consider direct mail; how many of your direct mail customers also receive the same marketing message via email? Which one do you consider most valuable? The DM piece that lands on their doorstep for immediate attention? The email that can be fully tracked from opening to clickthrough? How about using both mediums in tandem, holding back on emailing those that have received DM with a follow up email if they’ve not responded within a certain amount of time? Or perhaps you want to reduce print and distribution costs and encourage more customers to use their emails?

The key here is to think about your website as complimentary to traditonal marketing tools.

So how does this relate to online branding? What I’ve found time and time again is that brands are reaching a point now where they’re finding it difficult to continually grow the online channel, with 2 key elements missing from their websites; brand proposition and user experience.

Brand proposition is the portrayal of your core brand values; who are you? What do you do? Why should I buy from you? Why should I choose you over a competitor? How much? What’s in it for me? Etc etc

With brands driving traffic online through traditional marketing channels and seeing a heavy bias towards brand searches online, they’re not increasing their visibility online and not converting ‘brand virgins’ with their ‘buy from us if you want’ mentality. Sorry guys, this won’t cut the mustard too well these days, in a world of millions of choices and websites. Today, you need to maximise potential conversion of every visitor to your website to stand out against the competition.

User Experience covers your brand, product, marketing, pricing, competition and of course the look and feel of your website, ease of task completion and cohesion of each page to delivering a sale or lead. So much to cover, just where do you start?

You know your brand proposition right? You know your product(s) and you have a website. That’s a great place to start. Next, put yourself into the shoes of your customer and look at the various marketing messages you throw at, sorry deliver to your customers. Does online and offline carry a unified message? Do the designs and branding at least look like they’re from the same company? The place to start is to see things from the perspective of a new customer. Do you ‘get’ the brand and would you have enough information to at least visit the website?

Marketing material is as much a hook into your website as it is a communication channel. That’s why the user experience spans each customer touchpoint before you even consider how well your web delivers the user experience.

There are lots of optimisation techniques to hook potential customers in, but i’l leave that for another article. The next stage is how to give the customer what they want when they visit your website. Now worth pointing out is to appreciate and understand 2 things.

1) your customers expect to be able to find what they want quickly. They are the most important person on your website at that moment in time.

2) your customers don’t always know what they want!

Kinda tough to please everyone? Using a basic understanding of the hunter, tracker, explorer model of user journeys, you should at least cover off 80% of you user’s needs. For the other 20% ensure they can call or email you with ease.

Landing page optimisation is also in itself a key area of study, with numerous techniques and tools at your disposal. However, don’t rely on a one-size fits all design and consider instead at the very least, creating 2 landing pages for your marketing; for new and for repeat visitors. Their likely needs will vary with a number of areas of crossover.

Note: whilst this may not be the case for all businesses, invariably customers do not always type forward slash web addresses in. They simply go directly to the homepage. So if your DM campaign directs people to a dedicated microsite, you may be missing a trick if it is not also accessible via the homepage (an obvious challenge for below the line marketing).

Considering you have created 2 basic journeys into your website, again ask yourself whether you really fulfill the majority of customer needs. This is where an internal brainstorm with your brand/product team or external focus group with potential customers can really make a difference in uncovering hidden gems.

Creating a compelling brand journey doesn’t require a completely new approach to the layout of your website. Instead, incorporating existing collatorol into your key product and sales pages can in itself make a huge difference.

The point I wish to stress is to assume nothing. Once you crack it, your website will not only serve as a fully functional tool for those that already know your brand, but will also become a central platform in driving new business through your brand marketing.

What brand marketing challenges do you face? Get in touch for some advice or feedback.