CX & Design May 26, 2017
Dynamic content on websites: five problems (and solutions)
The implementation of personalisation is quite extensive in outbound marketing, but it is not yet widely applied on websites. Despite the fact that you can target customers much more accurately by combining data points.
I expect that all websites with high conversion goals and/or an eCommerce approach will be 70 percent dynamic within three years. After all, we are getting better at finding out what kind of user is visiting a website and what his or her characteristics (via data points) are. This means that you can evaluate the reason for a prospect’s visit from their very first visit, and will be able to target him or her better for conversion.
The search alone provides a lot of information, which means you can show more dynamic content. This is something our customer Vacanceselect is already taking advantage of. Say, somebody types “campsites in Tuscany with swimming pool” into the search engine. They will be shown a page with the header “campsites in Tuscany with swimming pool” and a matching picture plus text. Now, if the customer looks for “forest campsites in Tuscany”, they will be shown the same page, but the content will focus on the nearby forest. The images are linked to the search term using Google’s new Artificial Intelligence technology, which has a 93 percent accuracy.
You can take the personalisation of websites even further by linking customer information and click behaviour information from the CRM to offline store visits. You can achieve this through a loyalty programme or by identifying payments with a hash that is also attached online. The use of beacons is also becoming increasingly popular, allowing you to retrieve email addresses from visitors and match these with your database. Based on the profile information in your marketing automation tool, you can create groups of specific audiences. 1st and 3rd party cookies via your DMP complete the picture. Once you have combined all this information, you can make most of your website dynamic. Of course, there are also fixed components, such as contact details, which is why we say it is possible to achieve around 70 percent of dynamic content.
Once you have the necessary data, you need to make strategic decisions about what you show in terms of product information. Is the relevant person in the orientation or the decision stage, for example? Harder data comes in handy here too, such as demographic characteristics and which product category was selected.
What is stopping you?
Believe it or not, we are not presenting a difficult-to-imagine, futuristic scenario here. It is all a matter of linking sources. So what is stopping companies from getting started with on-site dynamic content? Here are a number of arguments we hear often, plus our responses:
“The customer is looking for recognisability and predictability.”
The fact that personalisation increases conversion has already been proven. That is why it is widely used in email marketing. There are no figures available for dynamic websites yet, but it makes sense to assume that the effect of personalisation will be similar and that it will not cause alienation.
“It is impossible to manage this much content.”
There is specific technology for managing content available nowadays. In addition, you can monitor the quality by comparing the dynamic site’s performance to a test group that will see the default pages. You do need good content, however, which is something that is often underestimated. The more customer journeys are taking place, the more content you need. Not just in the shape of product sheets but also through the use of good visuals.
“The technology just isn’t there yet.”
If you are working with a legacy system that does not allow for dynamic websites then it is a matter of budget, not technology, because it is definitely available. Out of financial considerations, investing in a DMP may not be for everyone, but a decent CRM system and a click tracking tool should be feasible for most. Additionally, more and more CMSs and commercial platforms offer an option for dynamic content, although this is based on what the customer has clicked without including other data sources. So this means you are missing out on opportunities for hyper-personalisation.
“It requires significant investment, and it is not yet used on a large scale,
so it is difficult to make a business case.”
This is a valid argument. But … if you don’t shoot, you will never score (J. Cruyff). There are many sites that already offer predictive recommendations, such as bol.com, Wehkamp and pretty much all other major eCommerce players. However, this is only a limited application of dynamic content, predominantly based on the clicked product category. It is not possible to realise seventy percent dynamic content this way, but you can use it to measure whether personalisation works by comparing the conversion achieved by these sites with a control group that was shown the default website.
“We do not have enough of the data points required to be able to deduce where the customer is in the customer journey. So it is not possible to create a site that is personalised enough to achieve conversion.”
You can initially start personalising your site based on your own data. Of course, the more extensive the customer profile, the better, but third party data is no hard requirement. You can begin on a smaller scale, for example by showing personalised recommendations and expanding the dynamic content if this turns out to be successful.
On the way to dynamic content
For companies that get past their cold feet, the first step towards a dynamic website is looking at how many data points they have, and to what extent they can personalise their website. It is of course important that the customer journeys have been charted, and that the company knows which target group displays what behaviour. Then it is just a matter of processing all these data points so all the different data can start working together. This does require a unified way of storing data, otherwise it will not work. The location where the data is stored is also important: it should be easy to retrieve and visualise and it should be simple to run reports. You look for patterns that you recognise yourself, or get the AI to analyse the data for you. Of course, the AI will need to be fed and trained by people, so it is not an easy way out, but, thanks to its self-learning character, this technology ultimately saves a significant amount of time.
Companies that want to seriously start working with dynamic content on their site do have to deal with some obstacles. It goes without saying that things will only work when you use high-quality data. After all, you need to make sure you do not accidentally address the wrong person on your site. When people share a computer, profiles soon start to overlap. An intelligent algorithm can distinguish between these users. And if there is any doubt, you just ensure you display generic blocks of content.
Furthermore, the algorithm needs to be tested thoroughly before you switch the site to dynamic content. Making mistakes is inevitable when using an AI, because the algorithm is fed by people. But that is a stage you will just have to work through.
Another issue is the effect dynamic content may have on the load time of the website. Good news: it barely makes a difference. Calling large numbers of scripts in real time hardly takes any time at all nowadays. Sending a request to a page and returning the content occurs within a millisecond, so it has no effect on the load time of your website. I also do not expect it to adversely affect the site’s position in the Google ranking. The content presented to the visitor is basically the same, it is just packaged differently. In addition, Google tends to consider a good user experience to be a positive thing. In fact, it is not unlikely that dynamic content on websites will have a positive impact on the SEO, as is currently the case with AMP sites, for instance.