The definitive guide to Warming up IP’s by Todd Fraser
Learning how to look after your IP’s can be a daunting task. The main rule is you must treat them like you would your own children, do not take your eyes off of them and give them lots of small bits of information over a long period of time to help them grow.
In order to give you a clearer idea of starting from scratch or bringing an IP ‘back from the dead’ I will describe a recent situation that happened to me around 3 months ago.
I send around 10 million legitimate CAN-SPAM complient emails per month. This often spreads over 5-8 new campaigns per week.
Around 3 months ago all 10 of my IP’s all had a good reputation according to senderscore.org and senderbase.org. I was sending around 10-12.5k an hour per IP and delivery was a consistent 97-99.8%.
All of a sudden delivery’s halved and reputations disappeared, Panic.
The first thing I did was look at the delivery results. Around 80% of the rejections were ‘due to spam like characteristics’. This was strange as spamassassin.apache.org gave me a spam count of 0.0 (the lowest possible) and I am extensive in testing and building high standard best practise emails. Something wasn’t right. The remaining 20% of rejections where hard bounces (mainly dead email accounts). To avoid any future excess hard-bounces I decided it was long over-due to get our data cleansed. Using dbg.co.uk and their advanced tool called ‘Verify’ I ran the data through its extensive cleansing system and removed any ‘foreign, suspicious, rude and duplicate emails’. At least this gave me peace of mind with the data to get to the bottom of the problem.
The second step was research.
By reading a variety of blogs such as blog.wordtothewise.com and dmaemailblog.com and combining knowledge obtained from reporting websites such as returnpath.biz I was able to find out that there was an even bigger than normal issue with getting into hotmail inboxes. Due to the beginning of its refurbishment, rules were getting stricter.
I couldn’t improve on my emails, my data was now clean and the only option was to completely start again on building up my IP reputations. What a daunting thought knowing volumes would become dramatically decreased but my expectancy to send 10 million emails wouldn’t!
After informing colleagues of the news I then began calculating and testing everyday over the period of a week Monday-Sunday.
Testing showed me that all IP’s would need to start off sending consistently at a rate of 500 hotmail emails per hour 24 hours a day (or as near to this as possible). This meant for all future sends I would have to exclude hotmail and then add on 500 hotmail only emails per hour to the total send. For example a 24 hour send would amount to a mere 12,000 hotmail emails. Luckily the ‘non hotmail’ email volumes could be ramped up to 2,000 per hour therefore providing me with 48,000 emails per day per IP amounting to 480,000 emails per day of non hotmail email addresses. Unfortunately open rates suffered due to the erratic sending times but at least delivery rates were a consistent 96-99%.
After a few days of consistency it was time to test out a new volume increase. Hotmail was up to 1,000 per hour and non hotmail was ramped up to 4,000, this was successful. After a couple of days I decided to push it and increase hotmail to 4k as well. This proved a bad decision and delivery rates plummeted to around 60-70%. I then went back to 1,000 and started again. After leaving things a week I went up to 2,000 and this steadily continued for the forth coming weeks.
The main key to take away here is that it DOES take time, sometimes weeks. The key is small volume increases and consistent delivery over a long period of time. Hotmail would rather see consistent small volumes of email rather than huge volumes over a few hours.
Eventually I was able to take the risk of merging the two sends and hotmail no longer needed to be treated like a separate entity.
Final thoughts
How was I able to avoid conflict with clients/colleagues?
By doing extensive research I was able to provide facts and proof of the situation. This allowed me to explain fully to my colleagues what needed to be done. By reiterating the importance of this practise, my colleagues were able to warn clients of longer send times and this kept aggravation to a minimum for all parties concerned.
I hope this has given you more of an insight into IP’s, how to maintain reputations and what to do in the event of a crisis such as the one above.
