Link Love Notes: What works in SEO outreach emails? #LinkLove
(Notes taken at Link Love London 2012 conference)
Mike King, the newly announced Director of Inbound Marketing at iAcquire, led a very informative and data-led presentation on linkbuilding and outreach emails.
Using data both from the internal team at iAquire – and their iRank system – and aggregated, anonymous data from the linkbuilding and outreach platform Buzzstream (which the Attacats are also currently using), Mike tried to answer the question that many linkbuilders have on making a successful outreach email:
Why can’t I get responses?
Mike’s process and methodology
Mike described their agency process for developing and refining their link prospect lists, incuding:
- Remove blue list pages (i.e. never gonna link, like Wikipedia)
- Remove black list
- Filter further by client requirements e.. minimum PR, industry…
- Manual review of link potentials – which ones to really reach out to?
This process can take them down from a full link prospect list of 13m pages to 1.5m, then to 100,000 pages, then down to the final list for manual email outreach.
tools to build your data on prospects
Mike then went through and described a few lovely tools that can be used to build up information on link prospects (some of which you’ll hopefully know of):
- Screaming Frog – use it to crawl your prospects in list mode
- Then use SEOTools Excel plugin to get any other metrics
- Then add any other Open Site Explorer data you might want
- Have a look at the SEOgadget Excel tool
He described how it’s important to consider additional value of pages/domains that are shared or scraped more than others (for example a post on authority site SEOmoz will be copied and shared all over the show).
SO, the study…
Within the emails that fit the criteria, they looked primarily at overall rate of response and overall rate of link acquisition or ‘closing’.
- dataset large enough
- conclusions made in a vacuum (looking at each metric individually)
- buzzstream has varying link building practices, philosophies etc.
- correlation is not causation
- iRank is controlled methodology, so controlled
the things that influence response rate
Does the gender of the person sending the outreach email have an effect?
Surprisingly to the audience women didn’t perform better – women get a higher response, but men a higher close rate. But overall it’s not much of factor.
So you don’t have to act like a woman anymore @timbarlow…teehee.
Personalised with their name gets much higher close rate. Custom, in that you just refer to the the domain or business in the email, do BADLY. Even general, non-specific inquiries do better.
*Use rapportive or RapLeaf to get more information on the prospect, but make sure you get the contact’s name!
‘hey’ did well but too low sample set
Hi did best (better than hello)
day of the week
On buzzstream (emails sent every day) shows less email on weekend, but SAT & SUN (& FRI) got best results. Boomerang for GMail can schedule.
time of day
9am best response rate from sending from their system, but buzzstream has best response rates in the middle of the night.
Basically people look at their email in the morning and pick them up in the morning. So send at night!
number of emails
Generally low after 4 emails (because it dies, but will jump back after loads). SO BE PERSISTENT!
Longer emails have better close rate, but short had better response rate. (1,000 characters + is long)
first touch via twitter
If tweeted first got much better response. If using Buzzstream ensure you connect your Twitter account! 35% got response if started with twitter.
how do trust signals perform?
e.g. phone number in email, logo.
Having no phone number in email actually performs better, got almost 5% more response and 2.5% more link closures. Maybe people think of it mentally as spam if it has that?
Logo in email performs much better 6.86% better. Linked logos have the highest rate of closure (10.5%)
Mike makes new twitter personas (still his name though) for different niches and clients.
the worth of a link – via mike essex
He says a link has ‘link equity’, as you need to draw it back to the monetary value of the link.
(Link brokers charge $100 per month for PR5 link) x (no. of links monthly x price of labour for those links monthly) = the value of link equity you are building
What would it cost to achieve the same incrememtal traffic increase through paid search?