How can recommendation engines help with the various phases of classified ad sites? To kick off the new year, we’re back with part two of our series on classified sites.
The vision behind motivating existing users
Recommendations placed on any site can increase the quality of user experience and keep visitors occupied for a longer period of time. Apart from the instant advantages and hard numbers, what does that mean for a company in the long-term?
Giving a smooth, exciting experience for the visitor on your site is essential. They should see that your offers always fit their actual taste and needs. As we move into 2014, this accurate targeting is a mandatory requirement. In addition to that, you can be the lifetime partner of your users through different lifecycle phases. You can react to their changing interests – moreover, to some extent, you can also generate those interests…
Just one example about the mentioned lifecycle phases: if you can see that your visitor was browsing baby strollers and cribs almost a year ago, you can be quite sure they will be interested in baby feeding chairs and baby dental products by now. This demonstrates that there is added value in knowing your customer’s history, which you can translate into revenue! In this way, you can be someone’s favorite shopping assistant without even meeting them!
Momentarily active users
How can you optimize the outcome for people who are on the site right now?
If you enhance the content experience on as many areas of your site as possible, you increase the satisfaction of online visitors and boost click through rates. The rise in CTRs varies according to the placement and appearance of the personalized recommendations.
Our experiences show that the best places to put the “recommended for you” boxes are ad detail pages, after ad-reply pages, result pages and zero results pages. Every attribute of the personalized ad box has an effect on the actual CTRs. If you use the wrong words (“Recommended for you” or “Take a look at these items, too?” etc), the wrong colors or – as mentioned above – the wrong placement, your results will be less satisfying.
Note: as a client of a personalization company, you should have full control and the final decision concerning the placement and attributes of these boxes.It is crucial to A/B test every mentioned aspect and evaluate the results together with your chosen personalization expert. Although there are no two businesses completely equal to each other, in the coming weeks we will give you a more detailed compilation of cases that proved to be successful at our classified clients.
Let’s see the logic the recommendation engine follows on the different types of pages. On the main page, most popular items are usually shown to every new visitor. The more the engine knows about the user, the more personalized offers it can give, so the “most popular” section will gradually turn into a mixture of personalized content and trending items.
On result pages, we usually apply a personalized approach based on user browsing history. If the visitor uses keywords in the search box, then we take those keywords into account too.
The approach above is also effective for the zero result page, when the search didn’t yield any result. We can still provide “next best options” based on browsing history and the keywords applied.
On ad detail pages, we select items that users, with similar tastes to the actual visitor, clicked on (collaborative item-to-item logic). These are then mixed with products that have similar attributes (content-based item-to-item logic).
The personalization of exit pages is one of the most important elements in prolonging the time users spend on a site. Every page that is likely to generate a noticeable amount of site abandonment is an exit page. For example, the very first page a visitor sees when entering a site is an exit page in itself. If your visitors don’t find anything engaging on the main page they will just leave your site right away. Other typical exit pages include zero results pages and after ad-reply pages.
Existing but momentarily inactive users
How can you bring back earlier visitors? With the help of cookies, we can identify visitors even when they are not on the site. This way, through partner sites or ad networks, we can provide personalized ads to the momentarily inactive visitors. This is a technique called personalized retargeting. The success of retargeting also depends on where they are placed. Some classified companies have reached 800% growth on average in CTR on retargeted ads, while others “only” saw a 400% increase.
If your retargeted ad is hidden in the corner of a site, users will just pass it by without even noticing it. But if the ad is placed somewhere in the upper part of your site and users do not need to scroll down, the results should be more promising. There are no golden rules for banners, sites or audiences, though an ad with a picture (the bigger, the better) usually converts more than ads with only textual information.
How can you make a visitor from someone who has never been on your site before? Though it’s a little tricky, “pretargeting” unknown people is also a possibility. Let’s take a real estate classified ads site as an example. Smartly, this site decides to advertise on a weather forecast site. They could target the visitors of the weather forecast site by identifying what city’s or region’s weather they have searched for. Then, in their ad box they will show ads of houses from the same area.
As a second example, let’s take a media company with a wide portfolio, which owns a classified site as well. It’s possible to link the user-identifying cookies together within the portfolio. This way, you can recognize the same user on different sites and “cross-recommend” items for them based on their actions. This an excellent tool both for promoting rising stars in your portfolio and also to keep your visitors within it with more and more links.
Now, tell us your opinion about these methods, we are keen to hear your experiences! In the next part of our classified series, we will concentrate on monetization. Stay with us!