The proper way to do French SEO in 2016

I have been doing foreign language SEO for quite some time and in my opinion it is the easiest to compete in since the international scene is too crowded. French SEO is something that not many of us are familiar with, maybe due to the fact that most information on how to do it properly is found on French SEO and IM forums and if you do not know the language, you can not easily follow the conversation. Also, while most tools come preconfigured or are made exclusively by and for English, you will need to adapt them for this language.

The point of this article is not to teach you how to do SEO, but rather to give you some pointers on how to adapt your strategies and methods for French. It is not  a short read since I am going through everything I learned and know about it. Also, I will not BS you with complicated strategies, imaginary case studies, “go white hat” BS, super secret tools, tricks and shortcuts that will rank you in no time. Everything here will follow the KISS (Keep it simple, stupid) principle. There is no need to over-complicate things.

What you will learn and get:

  • A overview of the market and the potential for making money using French SEO
  • How advanced are the Google web spam measures for French
  • Which ranking factors are in play
  • How I used to do it for the last 4 years
  • Why it is not working anymore
  • What to do in 2016 to rank sites in France and Canada
  • Which tools to use and how to configure them
  • Some general tips on how to optimize you landing pages and money sites

The biggest obstacle

If you are like me (not a native French speaker) you will have quite a few obstacles to cross. Understanding the words of the language is not enough, you need to know the context, culture and to be able to think like French people do. Every step from doing your keyword research onward has to be in sync with that. For example my first attempts were to translate current sites into French and go with that. Big mistake! You should not translate, you localize and adapt the content, preferably by using a native in the process.

The potential for making money using French SEO

Let us take a closer look at French. According to Wikipedia it is ranked on the 18th place with 80 million native speakers. French is the official language in France, Belgium, Canada, Luxembourg, Monaco, Switzerland and many countries in Africa, America and the Caribbean. All of the countries listed by name have high standards of living and reasonably high incomes which make them perfect as a target for your affiliate products, CPA offers and any other way of making money using SEO. You might already know this, but French CPA offers are among the highest paying ones out there. Without going into more details, we can conclude that it has a lot of potential. Keep in mind that most things that were converting like crazy in the US five years ago and are totally jaded for the English market, are fresh new for France and you might be the first one offering them to this audience.

The Google web spam team and why it matters

Just think of how popular a specific language is and you will find out how big the Google web spam team is. For example the three most used languages on the planet are: English, Spanish and Mandarin. If we assume that Google and China are not that close and that the Chinese market is dominated by native search engines like Baidu, we can safely assume that the teams assigned for English and Spanish are the biggest and as such the spam algorithms developed, work the best to keep people like you and me away from ranking our sites to the top or at least they try to. But is it the same for other “not so popular languages”? No! Even Google cannot extend its span that far that quickly. Languages that are not that popular or spoken by very few people in the world will always be easier to rank. Just try to rank something in Bulgarian, it is like taking a candy from a baby. I have seen cases where a simple GSA blast ranks the site to the top and keeps it there for many years, no Penguins, no Pandas nor any animal from the Google Zoo seem to be having any effect here like they do for English SERPs.

But unfortunately the case is not the same with French. Although I have ranked quite a lot of French sites over the years, the Google Zoo problem is more pronounced here. It is not the same as English, but I can say with a decent amount of certainty that Google has tried to address the French spam problem over the years and made some progress.

The ranking factors

On the other side to see what ranking factors are in play, ask yourself the following: Which ranking factors are not influenced by language? Things like time on site, HTTPS and low bounce rate pop into your mind. I will not be discussing them here since they are the same for any kind of SEO and are not different for French. Instead, I will try to cover as much as I can on the content based ranking factors, since they are the ones where the least is known. But just for reference here are some of the other conclusions I have reached after doing some statistics mambo jumbo:

  • Contrary to what Google is saying, backlinks are one of the most important ranking factors, the number of linking domains is directly related to the ranking success of a site. Sites that rank for French keywords tend to have more links from .fr or .ca domains (no surprise there). It is a French ecosystem so, if you are planing on ranking a site for this market, try to get some “local” links. There are also sites without any links from these domain zones which rank well, although this is not so common.
  • As for the site speed ranking factor, most of the sites that ranked on good positions for competitive keyword phrases were in the allowed margins, but as always, there are exceptions.
  • HTTPS did not seem to have that much of an influence since for most of the tested search terms, there was a small number of sites with SSL.
  • Schema markup is still something new on this market and apart from the markup implemented in the open source platforms and some bigger sites, it is used in less than 7% of the sites ranking in the top 100 SERPs. I can not say with certainty that it has some impact, but since it is not that hard to implement proper markup snippets on your sites, there is no reason not to do it.
  • SERP click-through rates matter, so optimizing your meta title is crucial, just try not to turn it into clickbait, cos bounce rates matter too (confirmed with a test, around 250 HUMAN clicks in one hour time managed to move the site 3 positions up, bots did not have any effects, bad proxies maybe).
  • Exact match domains and keyword stuffed domain names still perform quite good and not only in smaller niches.
  • Age of the page you are trying to rank. Although Google likes fresh content, it seems that for more competitive niches it prefers aged pages and favors them quite a bit.
  • The longer the URL the more G. hates it. Shorter URLs tend to rank better.

Doing the keyword research for French

Keyword research is the first step in identifying the search volume for a specific niche. Here is where it gets interesting. If you come from a English or other language and French is not your native, start by translating the keywords you already know and use them as seed to get new keyword ideas in Google Keyword Planner, once you have selected a few, expand them using the selection you made. It helps a lot if you have a native by you to give you suggestions and pointers on what to look for, cos simply thinking in French is so quite different from thinking in English. While in some cases you logic can align, it is important to put the cultural differences and mindset into consideration.

Selecting the right domain name

Depending on which French speaking country are you targeting, it is a good idea to get country code TLD like .fr, .be and .ca. My advice is not to go for exact match domains, but rather for brand names. Although there are exact match or domains stuffed with keywords ranking in the SERPs, the long term strategy is to aim for brand name domains. If you are unable to get a ccTLD, just go for any TLD. You can safely use keyword + brand in your domain to you stay on the safe side. For multi language and multi regional websites you can use any TLD, but you will need to set up the right international targeting for you localized versions.

Setting up the correct international targeting

If you are unfamiliar with the international targeting by language and country you can visit this help page first. Basically, it is a way of letting Google know that you are targeting a specific language and country using HTML markup. There are many ways to do it and they all depend on the specific case of your money site. If you have a single site targeting multiple languages and countries, please take a look here. I will do a larger post about this topic soon, so for now you will have to do the research yourself but here are some tips to get you started:

  • When you define the opening HTML tag set the language for French with <html lang="fr">, this way you will be targeting all French speaking users regardless of the country they are in.
  • If your site is targeting French speaking users in France only use <html lang="fr_FR"> or French speaking users in Canada <html lang="fr_CA">.  The first part is the language followed by the country code.

Which configuration you will use depends on your case, but take time to learn how to set these tags correctly.

The Content

As with any other language you need unique content to rank your sites. This applies not only for the on-site content, but the content you are using for building your links. The rule for duplicate content does not affect translated articles, so you can safely use translated English content for your site and building link, just make sure that it is unique enough.

On site content

Now, an interesting fact is that the average word count of successfully ranked pages has gone up. Depending on the niche, 1.000 words might not be enough anymore, so 1.500+ is the new minimum. As for the keyword percentages and all the techniques used till now, there are some interesting changes happening with Google in the last few years. Here are some observations on the top ranking sites for you to consider:

  • Very few of the top ranking results have the exact match keyword in the Title tag. Google even stopped highlighting it in the Titles of the results.
  • Contrary to popular belief the keyword is not even present in the Description tag. Most of the pages do not have it or if they do, it is not something that Google gives much value to.
  • The headline H1 tag and the presence of the keyword are related to better rankings but H1 tags with synonyms of the search term perform almost as good if not better. In H2, H3 and H3 tags the exact match keyword does not seem to affect the rankings significantly.
  • 80% of the results do not have a single exact match mention of the keyword in the content. Things like keyword density are most likely dead for good. RankBrain seems to work.

Link building / referring pages content

When it comes to T1 content, it seems that when there is an article, it is usually longer than the SEO standard of 500 words (* data gathered from the readability experiment described below). Topical relevance plays an important factor here, meaning links from pages with related content to yours tend to contribute more to your ranking success.

So how get French content in bulk for link building purposes:

  • Scrape content and spin it
  • Automated translation of English articles

How I used to do French SEO for the last 4 years and why it is not working anymore

My first attempts on ranking sites for France and Canada were basically following the same scheme I use for English sites, GSA, Web 2.0, PBNs and all the other bells and whistles related to the matter. I will not talk about creating PBNs, setting up GSA campaigns, etc. in this post, there is enough info on this. The process is basically the same as for any English SEO campaign you would do. Instead I will talk about the content used, which is the main difference.

I used three strategies, content wise, for my SEO work. Using spun French content, using spun English content with French keywords and using translated to French content. There is also a fourth one at the end, but as you will see it is not the most appealing one.

  1. Since I was a big fan of WordAI and used it successfully on my English SEO work for tier one content, I gave it a try for French which they claimed it worked as a part of their foreign language spinner. I trusted that their product will work the way it worked for English without checking the results, a big mistake. Never trust anything without checking it and double checking it even if it comes from a reputable company such as WordAI. It turned out that the spun content was not unique at all, the performance of the French spinner was way, way different what I was used on seeing from the English Turing spinner.
    My VA ended up pushing those poorly spun French articles over my whole network. Why was this bad? Imagine publishing like 10.000 posts which are nothing but scraped articles with a just a few words changed. It is duplicate content, the easiest footprint of bad SEO that Google can kick you for. Did I want that, hell no! Do you want to see this in your WMT account?
    Google Manual Action
    The case turned out to be almost the same with Spinner Chief and X-Spinner. The problem was simple, French was not a language where you can just change synonyms and get something that you can use. Even if there were some changed words in the articles, they were not making any sense and not only grammatically.
  2. The second strategy was to use spun English content with French keywords pointing to my money sites. This turned out to be producing far better results than my previous one.
  3. The third one involved automatic translation of English articles to French using translation APIs and SEO Content Machine. This one was working good too.
  4. Now for some more profitable niches I tried creating the whole tier one on custom human written French content and manually spun articles which was not cheap, but with an investment of $2.000 in content only I was getting much better results than the two strategies I mentioned earlier. What seemed to be the problem? First it was expensive and was affecting my profit margin, second it was a slow process, and third it was costing me a lot of time to manage all the writers and check if the content is unique enough. If you think about it is not only the 2K I invested but also the time I invested in doing it. This was not something I could scale up easily.

What things look like in 2016?

The first strategy I used with the spun French content is unusable, WordAI, Spinner Chief, X-Spinner and some other names I have tried over the years are still not usable for French SEO. I have not seen any change from what the situation was two years ago and what it is now.

The second strategy with the spun English content and the French anchors seems to be losing its power. As Google got smarter it now detects this, and while you can get pass it with some luck, it is not a strategy to depend on.

The only strategy that still works is the third one, but having unique French content is a problem and an expensive one. Not everyone can afford to invest such amounts if they are not sure they know what they’re doing.

So to summarize the situation here are the final results from 83 SEO campaigns I have done over a period of 4 years:

  • The cons of the first strategy are the following: WordAI and SC have no API for French, X-Spinner has it but all of them produce spun content which is not usable. Links created with this content are not indexed well and there is a huge risk of penalties. No pros so far.
  • The cons of the second strategy can be summarized to the following, Google knows what you are doing and hates it thus there is a very high risk of getting slammed. The pros are that the strategy is relatively cheap to execute.
  • The cons of the third one are that you need to manually rotate the Bing API accounts and you are limited to articles without spintax, so each translated article can be used only once. Also using this content with GSA SER is not that easy and it is quite limiting. The pros are that it produces good results.
  • The pros of the fourth are that it works better than the three above and requires a lot less links to ranks a site. On the other side it is time consuming, needs a lot of work and huge investment in content, which are quite the cons if you do not have the time or the money to do it.

How to know if these results can be trusted? I have done 83 campaigns for 37 sites, 342 tracked keywords,  171 in the top 3 results for France, Canada and other French speaking countries, resulting in over 5.1M unique visitors in four years. In my book this counts as enough data to have a decent opinion on the matter.

Is there hope for automated content creation?

Having this blog brings many benefits and among the best are that I always get to try new tools before they are even released and sometimes even get them for free. Last November, I was invited to try a new French spinner. It was not released, it did not have a site, it was just an API on a IP address that simulated other well known spinner via a small modification in the hosts file on my PC. The results were the same as with the rest of the spinners that supported French and the only difference was that it had a bit more synonyms in the thesaurus. It was not exactly what I hoped for but the developers seemed super enthusiastic in improving their product. Months later they released The French Article Spinner “version 1 beta” officially and got integrated into SEO CM and although it had its imperfections it was better than everyone else at least uniqueness wise.

In March they released the v2 with the “Grammar AI” as they call it which does all the spinning according to French grammar rules and fixes the problems with the first version. I am using the unique SEO version or “unique setting” since. I have tried the “readable” and while it produces great readable content it is painfully slow.

So what are the results from the last few months testing this article spinner? I think it is too early to say anything conclusive but here are 3 things I am noticing for now:

  • The spun articles although not super readable are grammatically correct, 95%+ unique and pass Copyscape with flying colors.
  • I need less links on T1 to rank something;
  • Some of the spun articles are even ranking for some long tail keywords (quite surprised to see this);
  • More indexed backlinks without changing my tactics;

The only con if I can call it that, is now the developers decided not to sell the spinner to everyone (they even removed the subscription payments section from the site), saying that it will disrupt the French “ecosystem”. If you are lucky, you can get it and use it (for free if you provide feedback). I honestly understand this tactic and while there is a demand for such product, limiting its usage makes it more effective, cos eventually, if used enough, Google will figure out the patterns of the spinner like it has done with the other overused products on the market.

The readability score and its role in French SEO

We are all familiar with the fact that Google has integrated the readability score into its ranking factors. This is something that we have used in on-page optimization for a while.  The Flesch–Kincaid formula is used to calculate score but the problem is that it is developed specifically for English. Another variant by Kandel & Moles, which is a modified Flesch–Kincaid readability formula can be applied for French.

Reading Ease = 209 – (0.68 * ( Syllables / Words)) – (1.15 * ( words / Sentences))

The experiment

I did a little research on 30 French keywords where I scraped the top 100 results for each keyword with the readability score (modified formula), Moz DA, PA, Domain, URL rank and Referring domains from Ahrefs. While the readability distribution averages varied slightly in each segment of next 10 SERP results, a relation can be observed where most of the top ranking sites had a relatively easy to read content and the bottom ranked had worse scores. While readability is not the only ranking factor, there was something to be explored here.

Another thing was that the average score was niche specific. For example, baby products related keyword results had easier to read content and other more technical niche keyword were on the opposite. This makes sense, if you target users are baby parents the complexity of the content will be much lower than if you target audience were rocket scientists. Depending on the profile of the audience you will have to adapt your readability score to a certain “industry average”, you can not write simple non technical content for users searching for how to assemble a Raspberry Pi or fix an issue with their PC motherboard.

The next experiment was to check if the backlinks reported by Ahrefs for each of the ranking URLs had the same logic. The result was that sites that ranked higher had links from pages where the readability score was very close to the one of the page, which was in the niche average. What does this tell us? While this case study was for French SERPs and it is also valid for English, content matters not only for your on-page optimization, but also for the pages where you get your backlinks from.

The confirmation test

Once I had this data to work with, I started to test the theory on small affiliate site which ranked 6th. The niche was not so popular and lets say the situation has not changed at all for the last 9 months. The same sites were on the site positions, no new players no changes in ranking positions.

Since there were only 105 links pointing to it and most of the other competitors were having the same backlink power, I decided to rework the content on the linking pages to be more readable. Reworking 105 articles was too much work so I settled up for 1/3rd of them.  Five weeks since the change I got some results, with the new changed articles being naturally re-indexed by Google, my ranking improved by 3 positions. The same 10 results on P1, but now I was 3 positions higher. When I weight the investment for the test and the increased number of user and conversions I broke even in 4 months. So in the long run this was a smart decision.

What tools do I use?

GSA Search Engine Ranker

While it is primary built for English SEO, GSA SER can be modified to fit our French link building techniques. Although some would argue, I am still using GSA to build tier 1 links, but in moderation. It is always good to have some link diversity and I am talking about links form different types of sites. If you setup your campaigns right, creating links on sites with related content to yours, the results are quite good. Now, you are all familiar with the “Generic Anchor Text”, but how to use it since 99% of the anchors are in English and not only that thousands of GSA installs running all over the world are spamming like crazy with those same 800 “generic” anchors. I am sure that most of you have never looked into what exactly these anchors are. My advice here is if you do not have the time or you do not plan to do much French SEO, simply do not use them. On the other side if you are like me, spend some time creating your own French generic anchors and modify the generic_anchor_text.dat file in the GSA folder. If you are asking me for tips how to build that list, I built mine by scraping over 450 sites and creating frequency tables for the anchors which were than filtered by hand.

On the project options tab, there are 3 settings you might want to configure. If you are using search engines to find targets, right click over the “Search engines to use” field, click “Check none”, “Check by language” > “French”. If you want to create links from site with French content you might want to take a closer look at “Skip sites from the following countries” and “Skip sites from the following languages”.

Just remember use GSA for T1 in moderation, for T2/T3 you build as many links as you want.

FCS Networker

For FCS Networker there are 2 important things to remember. If you are going to post French articles, posts or anything else, don’t use spintax, only already spun articles. The problems with the character encoding are still not fixed. Also you might notice that almost all PDF submissions exhibit the same issue. Recently I am seeing some poor performance from this tool but I think it is related to the public proxies they are using for posting to the sites.

RankWyz

I have not been using RankWyz any French SEO work (I got it only recently), so as soon as I have some observations on the matter I will update the post.

SEO Content Machine

For all content creation tasks, except my money sites, I use SEO Content Machine. When creating a new task on the “Article sources” tab,  from the “Scrape content” panel select via Google France. Also it is a good idea to add some French sites related to your topic in the “Custom sites” list. If you decide to add links to your articles, do not use any generic (same as GSA). For the link position in the article, I have found that “Replace keyword or insert randomly” and “Insert link contextually” work best with foreign content.

There is another important feature in SEO CM that you might want to try, the bulk article translation tool, which is also quite good, but you will need a lot of API keys for the Bing translation API. Unfortunately there is no way to use more than one at a time, so try to fit your tasks into the API limits.

The French Article Spinner

Now for the article spinner part, I personally prefer The French Article Spinner, since it creates the most unique French content from all the others I have tested. Since at the moment it is not commercially available (you can get a free beta access on their site if you are interested) you can use what you already have, but as I have previously discussed the other spinners you should be aware that the results will not be very good.

Proprietary tools

So far I have not found any software that creates and manages web 2.0 properties the way I like, I have developed account creators and submission tools of my own. Don’t ask me to share these tools, they do the same job as FCS Networker and RankWyz but in a different way, so for now I have decided not to sell or distribute them in any way.

A few final tips

Since you will be targeting France and most French speaking countries, you money site will be in French. One of the most important tips I can give you is that the quality of the content in terms of proper language, phrases, grammar and punctuation can make a huge difference in conversion rate. While for International English landing pages you can get by with some mistakes and not that great English, the case is not the same with French users. They seem to be valuing their language quite a lot and even a small inconsistency in your translation can turn them away from converting.

Spend a few bucks more on quality translations for your landing pages and get a native to proofread them. I have doubled my CR after implementing some very minor changes in the texts which were basically discrepancies caused by the translation from English to French.