Link Building Case Study – Results of 515 Outreach Emails

We sent five hundred and fifteen emails asking for backlinks meaning, and today I’m going to share the stats from our campaign, the exact process we used in key learning points that you can use in your link building efforts. Stay tuned. What’s up, everyone, Salmo here with us, and this is the second episode of our three part series on creating linkable content and actually getting links to it. So if you’re new to the series, the first video was all about strategically building a statistics page that we thought would generate links. And even before we started creating that content, we found roughly 40 to hundred websites we can contact. Of course, this number is unfiltered and unvetted. And since I showed you how we created that content in part one, now I’ll show you how we built links to it here in part two. So as we were planning our experiment, we wanted to report accurate results for our link building campaign. So to isolate backlinks from our email outreach efforts, we didn’t promote the post to our existing audience. So no email subscribers, customers or social followers. On top of that, we hid it from our blog index and category pages, RSS feeds, and we even know index the page so it couldn’t be discovered and search. The post was basically invisible unless we sent you an email. So without further ado, here are the results from our campaign, we sent five hundred and fifteen emails, four hundred and seventy three were deliverable and forty two bounced. And based on deliverable emails, we had a reply rate of seventeen point five five percent in a conversion rate of five point seventy five percent, which works out to twenty seven lengths from unique websites. But to other cool things happened.

Number one, we got an additional five referring domains from sites we didn’t reach out to. And the most probable reason is that some people were discovering our page from those that link to us and from the odd social shares in total. We had around nineteen shares on Facebook and two on Twitter, none that we started. So that brings our referring domains count to thirty two or six point eight percent conversion rate. And number two, some of the sites we reached out to link to us more than once, both from new and old posts. So in total we had thirty six editorial links from unique posts. But let’s use our first conversion rate of five point seventy five percent on referring domains as I think it’s a better reflection of our results. Finally, I want to give you an idea of the quality of links we picked up. All the links we built were editorial links on contextually relevant pages. Now, as for metrics, let’s simplify it and talk about just two domain rating, which represents the overall strength of a website’s background profile and domain level search traffic. So this is what the domain rating distribution look like for our acquired links. Nine of the thirty two referring domains had a D-R of 70 or greater. That’s twenty eight point one percent of our total referring domains. 12 of the referring domains had a d.r of forty to sixty nine, which is thirty seven point five percent, and the remaining eleven had d.r values of four to thirty nine, which is thirty four point four percent. Now let’s look at the distribution of websites based on domain level search traffic, and these are based on search traffic numbers from HFS, which are usually underestimations. So they don’t include other traffic sources like direct referral and social. So six of the linking websites get over ten thousand monthly search visits, one that actually gets over a million organic visitors. Another six get between a thousand to ten thousand monthly organic visits. Sixteen websites get between one hundred to a thousand monthly organic search visits and forget zero search visitors, which all happen to be websites we didn’t reach out to. So are these results good, bad or average? Well, to answer that, you need to understand the full scope of the process, things we intentionally didn’t do, like automated follow ups and things we weren’t willing to do, like link exchanges. So throughout this video, I’ll share some of my insights, as well as the more granular stats that no one really talks about. But I’d love to know your thoughts based on your experience and what you know of our campaign so far. Do you think the results were good, bad or average? Let me know in the comments. Also, I will warn you that if you’re new to link building, some of the concepts and thought processes are a bit more advanced than our other building tutorials. So feel free to ask any questions in the comments and we’d be happy to answer. Let’s dig into how we executed this campaign. So the first step was to get a list of prospects and we started by looking at the top ranking pages for the statistics that had lots of back links. So when we were prospecting, we saw that these three pages cumulatively have over forty two hundred referring domains pointing at them. And we found another page that had around nine hundred and eighty referring domains. So we opened all four of their backend profiles and said explore and set the one link per domain filter since we don’t want to page the same site multiple times. The next thing we did was look for reasons why people were linking to the stats pages. So looking at one of the backend profiles, you’ll see that the majority of links are happening because people are referencing a specific step. Pretty obvious considering the topic of the page. But which stats are being mentioned the most? Let’s head over to the anchors report to get a summarized view. As you can see, there are tons of links that can be attributed to this. Ninety three percent stat. So we continue looking through this report and jotted down around five to 10 popular stats based on the link anchors. So with this information, we needed to check off to checkboxes number one. Are there enough people linking because of these stats? If yes, then that could make a link building campaign worth the effort. And number two, can we justify a good reason for them to link to us instead of or in addition to the current link? So we went back to the back and support and search for mentions of the stack in the anchor and surrounding text of the link. And as you can see, there are well over seven hundred links from unique websites where they’re mentioning the stat. Ninety three percent. So that ticks off Checkbook’s number one. Now if we visit the page, you’ll see that the stat isn’t even mentioned on it. So that checks box number two. We have a good reason to contact them. So after analyzing all the stats, pages and their specific stats, we came up with two main backlinks angles, which were number one, the page linking two doesn’t mention this, that we have more recent data. And number two, the stat you’re mentioning is X years old, but new data suggests whatever the new data point was. So we exported the report and then we did the exact same thing with all of the other stats and all of the other stats pages. And because we were finding outreach angles for individual stats, this allowed us to easily segment our link prospects so we could send personalized emails to everyone in a somewhat automated way without being spamming. So we imported each of these reports into Google sheets and labeled each URL according to their segment. And by the end, we had a list of one thousand nine hundred and eighty six URLs. The next thing we did was an initial clean up job of our prospects list. So first I duplicated you or else from the same domain, since websites could potentially be linking to more than one of the stats pages, this was easy to do using Google Sheets. I just passed the domain from each URL, clicked on data and then removed duplicates.

Finally, I set out to analyze just our root domain column and then hit remove duplicates. The final thing I did here was to remove prospects based on Lync attributes in the pages as geometrics from our list of duplicated prospects. I further narrowed them down to just follow the links and websites that got at least a certain amount of organic traffic across their entire domain. And I pulled domain level traffic using the API with Google Sheets script editor. Alternatively, you can use batch analysis to just pasting up the two hundred domains and set the target mode to domain with all its subdomains. You can then export the file and run a V look up against the root domains. So after all was said and done, we had a list of nine hundred and two prospects that we need to manually vet. Now, if you’ve watched any of my other link building videos that I’m sure you know that I’m a huge advocate of using APIs and automation when possible. So to kick off our linked building campaign fast, I took all the URL and put them into a custom tool to scrape as many of the author names as I could. And in total I got seven hundred and forty one scraped author names, but it was far from perfect. But I still took all of those names and ran them through Hunter’s API to search for an email address. In total, Hunter sent back four hundred and fifty two email addresses and with email addresses that were found, I simultaneously random using Never Bounces API, which would tell us if the email addresses were deliverable. So after around 20 to 30 minutes of automation, we had a total of one hundred and sixty eight valid emails, ninety two capsules that needed some manual intervention in one hundred and seventy eight emails that were not deliverable. So these would need to be searched for manually. These simple automation has made our vetting process much faster because all we had to do for valid emails was check that the pages were of decent quality and they were still mentioning this, that we were about to pitch each prospect in the entire outreach process was handled by Vlad, who’s a member of our marketing team. He basically took a few hours each week to vet pages, find emails and send them now to simplify the vetting process. I sent a list of criteria for people we should and should not contact. Vlad would then set a specific status based on the page and domain so we could keep track of our campaign stats. So based on this list of statuses, we disqualified spammy pages, those in another language, since we wouldn’t be able to send effective outreach emails, ones where we couldn’t find an email and you URLs where we had existing relationships. And for all pages that matched our set criteria, we’d make them as good. The last thing we did in the preparation phase was write our outreach template and create personalization fields for each contact. Now, before we continue, I want to make a quick note on the concept of personalization. In my opinion, there’s a balance between sending a relevant, well-written email and sending a creepy one. First off, adding someone’s first name in your email doesn’t mean it’s a personalized email. That’s just the personalization field. But on the other extreme, you don’t want to say things that will be too personal, which might come off a bit strong. For example, saying, Kasem, then whatever your outreach pitches, by the way, I heard the weather in Toronto is twenty two degrees with a 90 percent chance of precipitation. Guess it would be better to stay indoors instead of playing golf. Love, Josh. While this might seem personalized, it’s a bit strange and also irrelevant to the purpose of your email. So the approach we took was to personalize emails based on a specific thing they said in their post. And of course, that was the stat, which again is why I think segmenting Lync prospects on large profiles makes a ton of sense. So here’s what our exact email template look like in pitch box. And since it’s basically a bunch of merge fields, I’ll fill them in as I read you a sample. Hi, Sam. I saw you mentioned how ninety three percent of online experiences begin with a search engine on your page about e commerce. So that’s that is actually 14 years old. More recent research from twenty nineteen suggests that this number has gone down to sixty eight percent. I think it’s lower because social and other sources now account for around one third of traffic. We publish this and a few other fresh stats here. Not sure if you’re actively editing posts, but might be worth an update if you are no pressure. Cheers, Vlad. Now, to set up these March fields, we basically created another worksheet called Outreach Emails, and it looks like this. And then within our export sheet, we basically just did lookups based on the segment, the contact Falzon, which would then automatically populate our merch fields. Then it was just a matter of exporting the sheet, uploading it to pitch box and then making sure that our emails made sense before actually sending them. Now, if you don’t use Pitchfork’s, you can do the same thing in pretty much any other outreach tool. And that’s pretty much it in terms of the entire process for setting up our link building campaign. Now let’s revisit that thought again about whether our results were good, bad or average. And there isn’t a straight answer because your mileage will vary based on way too many factors.

But for simplicity sake, let’s compare our reply rates to a couple industry studies. So our response rate of seventeen point five five percent was pretty high in comparison to the eight point five percent that was studied by Pitchfork’s and back Linko. So in total we had eighty three people actually respond to our initial email. Now with a good number of responses, our link acquisition rate on referring domains was five point seventy five percent, which seems pretty average at face value. So why the big discrepancy? Two reasons. Number one, we had a ton of requests for link exchanges and other things. So in total, eight people requested a link exchange. Three people asked us to pay for the link and six people requested something else, like a free account or for us to share their content on social media in these requests came from small websites with low authority and little traffic, as well as big sites with millions of visitors and high link authority. Now, linked exchanges aren’t something we participate in. We’ve never and will never pay for links, and we don’t just hand away Atrius accounts. So in that respect, we had very little power or actually no room for negotiation. But if we had done the link exchanges, traded HFS accounts and did the social shares, that would have broader referring domains account to forty six, which would have been a nine point seven percent link acquisition rate. And that’s pretty good in my books considering the number of emails we sent in the fact that we weren’t soliciting trades. The second reason for the discrepancy is that five people actually linked to the source of the stat rather than to the article we pitched. And these are people that actually replied to us. So there were probably more. And it comes with the nature of the type of the post. We created a curated list of the most recent stats we could find where obviously will be linking to the source on the page. And I’ll expand on this in the third part of this series. Another thing to consider when evaluating the success of our campaign is the quality of the links. As I’ve already mentioned, all of our links came from topically relevant pages and websites, and the majority of our links were from pretty good sources. Now let’s talk about some key takeaways as well as our shortcomings. So based on our overall stats, I think it’s telling about the current state of outreach for links. Based on my experience, the number of people looking for some kind of exchange or payment has skyrocketed. More and more people understand the value of a link and whether it be from knowledge or personal experience, they know that some people will give them tangibles for it. So in my opinion, it’s only going to get harder to get links which will require you to go with more of a targeted approach, especially since tons of people are sending horrible spam emails. So by sending good emails, you actually have a chance to stand out from the rest of the crowd. Now, there were two shortcomings for my campaign that I think contributed to a lower conversion rate. First is the age of the links. I’ve already told you that one of our pitch angles was that ninety three percent stat was fourteen years old and wasn’t even on the stats page. So there were definitely pages in there that had aged without an update in second, which was probably the biggest limiting factor to our link acquisition rate is that we didn’t send any automated follow up emails. Follow ups definitely help increase conversion rates. I’ve seen it in my own campaigns and the guys that authority hacker shared more convincing data from their study on six hundred thousand emails. They said that sending three follow up emails has at least doubled their results. So why didn’t we send them? Simple, we didn’t want to bother people without good reason. The main reason for this campaign was for a case study, not because we need tons of links. So the only people we followed up with are those that said they would link to us but didn’t make the change within two weeks, which, by the way, helped us get more links. So everything is out on the table. And you might be wondering if doing outreach for links is still worth the effort, considering lower than expected conversion rates? Of course it is. These stats are from just one campaign and our results can’t be generalized to all outreach email. Yes, you might find people wanting something in return these days, but it’s still possible to build good links without paying or trading for them. So to conclude part two, if you plan on doing outreach, I strongly recommend going with more of a segmented approach into stay away from sending ten thousand identical emails to every person. The additional effort can go a long way.

Now, outreach for links isn’t the only way to build links to your page. You can actually build them passively using PPC advertising. Or at least that was my theory. So in the final part of this series, I’m going to show you advertising campaigns that we ran to two different types of stats, pages and links that came as a result of our PPC advertising campaigns.

Leave a comment

Your email address will not be published. Required fields are marked *