Dovolenka.SME.sk is a travel agency. It arranges tours to foreign destinations for Slovak and Austrian travelers.
The client was dissatisfied with the conversion rate and portal traffic. He was preparing a complete redesign, and he needed to ensure their online presence so it didn’t drop on Google. And then, to make the most of SEO.
After the redesign, Google traffic increased by 13 percent. After stagnation in 2015, search engine traffic grew by 30 percent in 2016. A year later, it rose by another 178 percent. This resulted in increase gross sales by EUR 1.2 million per year.
It began in 2014, when the project manager, Peter Kuna, was preparing a redesigned portal.
Our task was to design the site and the content so that if users are looking for anything in connection with booking a vacation, they will be able to easily find us on Google, and to include phrases like “Last minute Turkey all Inclusive up to 500 €“.
The search habits of prospective customers were determined by keyword analysis. From its outcomes, it was clear that to achieve the goal, we need to redesign the structure of the categories and change the way we generated the content.
The original categories on the web did not take into account the habits of people who were searching to buy a trip. For someone, the most important factor is price, but for another, it may be the date. Many searches for combinations of multiple criteria directly on Google.
In order to be able to provide relevant responses to searches, we changed the basic category tree:
Click on the image to see in full resolution.
And we also suggested adjusting searches for tours in the way so that searchable combinations are available:
Partial illustrative section of combinations. You can see in the picture that users are often searching for “Last Minute + Country Name” combinations. “Holiday + Term” is searched fewer times, but even that does not have negligible numbers.
These suggestions were aimed at creating new landing pages. Then we had to “fill in the pages with the content”.
We suggested creating thousands and thousands of new pages. Although, it did not make sense to write tailored content. Time was pushing us because a few months later the site was being planned for a redesign. Such investment was at the same time very risky. We’ve had more important things to work on (and we still have, yet in 2018 many categories do not have manually created content).
For each category and combination, we suggested what keywords to generate in the content.
We had pages and we had content. But how can Google (and users) click through thousands of new URLs?
Because of SEO, we proposed to rewrite the search form for tours, the main menu, and then adjust the linking from the homepage and categories. A draft of internal linking from the site of the hotel looked like this:
Simplified design. Along with it, the client also received an assignment where each of the links leads, and how it should be generated.
Proposals interfered with the entire site. They should influence UX and conversion rate. They requested plenty of developer’s time.
We had been discussing Peter’s clients on numerous occasions regarding their highest priorities. We also had to decide how to deploy changes to not make it worse, but on the contrary, improve the conversion rate.
We bent suggestions to make it possible to program them in a reasonable time and at a reasonable cost.
At first glance, you might notice that the upcoming change was enormous:
Content, graphics and URL editing was planned. This is very risky in terms of SEO, especially if it’s a site with hundreds of thousands of URLs.
An improper transition would reduce traffic from Google a few times. If this happens, and even if you have the impression that you have done everything right, a drop will come.
We wanted to redesign it as soon as possible before the summer season. However, unfortunately, as it often happens, it did not follow the idea. Peter bet that the conversion capability of the new site would be significantly better, and we decided to redesign it before the start of the season, even though it was about to start in a month and a half.
This was no longer fun, as about 60 percent of the site’s traffic was coming from Google, and we could not afford a bigger drop.
We needed to do everything to minimize the risk of the drop. We had been focusing on the ideas that Google:
- Knew how the URLs change
- Had no problems with the technical setup of the new site
- Knew we didn’t lower the content’s standards, and that we didn’t want to extinguish any pages visible on Google
- That we were creating new content pages
In order to achieve this, we monitored:
- Proper redirection of URLs where we decided to change them
- Minimal content editing
- Well-engineered technical SEO
- Monitoring after the site migration
During web migration, the SEO rule applies: Do not modify the URL, and if you have to do so, redirect the original URL 301 to the new one.
At first glance, it looks simple. However, when it touches hundreds of thousands of URLs, you can easily forget something. To avoid this, we focused on every type of page, from hotel sites, through reviews to pagination between categories. We always had a pattern “before” and “after”.
|Demonstration of hotel redirection||URL before||URL after|
The less radical the changes you make, the less likely you are to drop on Google. For this reason, after redesigning, we tried to make the content changes on the site only minimally, ideally not at all. From the beginning, we left the original subtitles, titles and as much text as possible.
A common mistake is also to remove some content pages completely. On the new site “there is no place to use them” or no one knows that such content exists. Our job was to ensure that the content remained relevant on the web in terms of SEO.
The technical SEO analysis was made easier by using our Spotibo tool (at that time, it was still an alpha version of the tool called Simulyzer). Thanks to partial automation, we were able to check every single URL.
Even if you’re testing the upcoming site thoroughly, only the sharp version of the site shows if there’s a problem somewhere.
We had been monitoring whether Google sent users to the error pages (404) due to a redirect error. We had also been monitoring whether any page type had lost positions or created indexing issues, and we tracked other important metrics.
Fortunately, we did not have to fix any unexpected problems, as the developers at Dovolenka SME did a great job.
With this approach, we achieved no traffic drop. On the contrary, we had seen a slight improvement:
After this “one-time” intense task, we backed off from the project for a while, and we only provided occasional consultations. We have devoted time to several other SME projects.
Overall, the year-to-year comparison before and after a redesign, in 2014, we gained 13 percent more traffic.
On the one hand, we succeeded in meeting the main goal of managing redesign without a break. However, since then, the half-year had passed and an increase of 13 percent was far from the potential of the adjustments.
Something was wrong.
In mid-September, we introduced Peter to a new SEO strategy for the next year, and we wanted to spend at least 12 hours per month on SEO. We planned to focus, for example, on:
- Link building – getting fewer links from top-quality sites. We had been planning to use the Cestovanie.SME.sk project for support. We completely snubbed phrases where most of the competition had been already (for example, “last-minute Spain” and similar).
- Linking the project from other sites to SME.sk
- SEO monitoring – if some problems arise to know about them straight from the embryo.
After approval, we started working. Shortly, we encountered a problem that could have hindered the growth of Dovolenka SME.
After the redesign, Google had indexed hundreds of thousands of unnecessary URLs. It indexed crazy combinations of categories and hotel offers. Where was the problem? Each hotel’s page was dynamic. Based on the user’s needs (city of departure, date, …), the website generated a tailor-made offer.
The URL of this offer contains many parameters. It looks like this.
One hotel could generate hundreds of URLs with very similar content. Because of this we had totally broken internal linking. The reference value had faded away into insignificant pages, and the important pages received much less of it than they should had.
Even though we used a canonical tag with which we informed Google about duplicate or similar content, Google ignored it.
We had known earlier that the ideal solution would be to separate parameters with a grid character (#) to make Google ignore excess parameters. However, this adjustment was technically very demanding.
In any case, however demanding it was, the programmers had to start solving it. Gradually, we received toxic URLs from internal linking and changed them to grids. But, it was very slow.
The technology that was used to develop the project did not make it easy to change the question mark behind the grid. Maybe it sounds trivial. However, to get it working, developers had to adjust the framework built on the PHP programming language. They had to “bend” it for their own needs.
The whole problem was managed (almost) completely to solve a year later.
Important pages had great losses due to broken indexing. Although pages with a new structure were serving customers, and in the meantime, we had also been working on improving content and link building, the traffic had not been developing for the better.
Sometime around the end of summer 2014, Tripadvisor launched its Slovak mutation. And in 2015, Tripadvisor began to dominate in the search of hotels. Even the Slovak version of another foreign giant, Booking.com, had risen in search of hotels.
Together with technical problems, traffic coming to the hotel’s pages for Dovolenka SME had significantly dropped. In the summer of 2015, it was more than 40 percent:
As if that is not enough, the demand for tours fell in 2015, at least in the countries where the site Dovolenka SME dominated on Google. For example, last minute interest in Turkey fell by almost by half:
Despite regular SEO work, year-to-year traffic from Google fell by 21 percent. However, thanks to a very well-managed UX, the gross sales from SEO was higher. Even more than doubled (according to the attribution model of last-click).
To the next year, we went with a tidy web and better internal linking. We mainly focused on the development of information content and the collecting of backlinks.
For every long-term project, we were also continuously analyzing the competition. In the case of Dovolenka SME, we used it a little untraditionally.
Invia, the largest travel agency in Slovakia, had an internal search on the web, and the results were accessible on Google. So if someone searched for something, Google indexed it. In this way, they covered many long-tail phrases without much work.
However, the disadvantage of this solution is that user searches were publicly available on Invia. We just needed to “Google” those results.
Thanks to the keywords we’d acquired, we created content that had earned us about 60,000 visits in 2016 in 2017, and up to 200,000 visits in 2017.
Resolution of the indexing issues was not enough for us. We worked on the content of the hotels, their internal linking, and sometimes we received some backlinks.
The year-on-year comparison of hotel’s traffic was as follows:
Not only did we manage to reverse the situation from last year, but compared to 2014 before the downfall, hotel’s traffic rose by 30 percent.
And paradoxically, the interest in Turkey trips that was key to us was even lower than a year ago:
We also had a significant drop in pages with trips to Turkey – 37 percent. However, we had worked to optimize many other countries, so we did not feel as much of the decline of Turkey.
We managed to reverse the negative trend and started a rapid growth at the beginning of the summer season. We will not compare it with the unsuccessful year 2015, but with 2014. The traffic was up 30 percent. Months at the end of the year showed growths of 170 percent or more, indicating that 2017 could be extremely successful.
During this year, we continued with the working strategy. We mainly focused on monitoring the technical state of the site. We only occasionally consulted with Peter and other members of the Dovolenka team regarding on-page SEO.
When we were doing the whole year’s evaluation, I was instantly shocked. I thought we had an error in Google Analytics measurements. Thus, in the year-to-year comparison, Google traffic grew by 178 percent. And we do not mean a growth of 100 visits to 278. Overall, the traffic was higher by hundreds of thousands.
And gross sales from traded tours increased by almost €1,200,000 (last click attribution), thanks to SEO. The year-on-year comparison was:
This project has taught us patience and the effort to work, even if the project does not immediately develop according to the idea.
We should have tried to promote long-term cooperation right from the beginning, and not to leave the project living after the redesign.
Thanks to the experience on Dovolenka SME, we have started to do more things differently:
- When we have the impression that we see an important problem, we are pushing a lot more to solve it. We try not to disassemble the client with less important tasks.
- We pay more attention to the continuous monitoring of SEO; we monitor technical changes on the website.
- When planning, we look at trends, including how the market evolves, not only in the client segment but also in SEO. So, we could use our experiences of what we are doing now, and also in five years.
Get in touch with me at firstname.lastname@example.org.