I have outlined a simple 15 minute exercise that will help take you from a “Master File†scenario to a clear, organized and flexible segmentation set that will help you monitor engagement levels, establish better communication planning, and combat deliverability issues. Answer these 3 questions without logging into your ESP or database platform. Think about your customer, your brand, your purchase cycle… be the marketer! 1. What are my acquisition sources? Prioritize your acquisition sources based on risk and data quality. Sources such as new point-of-sale and sweepstakes should be considered higher risk for deliverability issues and spam complaints. 2. What defines the most active shopper? Let’s start with your one time buyer. That’s a given. Where do we go next? Do you have frequent promotions for a varied product line? If so, use smaller number of purchase ranges within a shorter window of time to categorize your purchasers. Are your products seasonal, do they have a defined usage period before needing to be purchased again? Do you have a limited number of SKUs? If so, choose ranges that are broader and cover a more extended period of time. 3. What distinguishes disengaged vs. inactive? Many marketers will divide lists into active and inactive segments. It is important to also determine subscribers who are starting to disengage from your email program but have not become completely inactive. Factor in your brand, customer, purchase cycle, mailing frequency and seasonality when determining these thresholds. If your competition offers very similar products and price points making it easy for your customer to start shifting brand preference, disengagement campaigns should be aimed shortly after opens and clicks start to flat line. Leverage your ESP to build queries matching the results from the exercise above. Determine whether thresholds need to be adjusted. Try casting a wider or narrower net to see where your subscribers cluster. Make sure you are not fragmenting your lists too far. If these segments are mailed, always make sure subscribers are only mailed once and not duplicated across segments. Now that you have made it through the deep thought exercise, stack the data sets like the example below and consider what kind of program ideas you could implement to help increase engagement and drive additional revenue. With only a 15 minute investment, I’m sure that this exercise will help you step up your email marketing game. Exercises like this that help us take a step back and look at the bigger picture. http://blog.bronto.com/
Engagement makes it sound so commitment-y. Be the marketer? Really? I'm glad someone is there to tell me that... Not really. The type of email you send is more conducive to generating complaints moreso than the source of the data. I can't think of any person who has only bought ONE thing in their entire life and you never really know who is going to buy what and why...and why does this article keep shifting between the "customer contact" marketer and the dedicated email marketer. Make up your mind! You keep talking about targeting your emails but you fail to target your article. The problem with categorizing data and limiting the types of offers you send to a particular section of your list implies that you really believe people only have one interest or are willing to spend money on only one type of product. You will do better to approach email marketing as you would TV advertising - it's the shotgun approach but it works and has been working for a long time. Rather than segmenting your lists you should be segmenting the offers you are mailing...if you really want to segment in the first place. Yeah that was really deep man. After all that thinking I've concluded that disregarding everything stated in the article would improve my revenues. Finish it off with a cliché or two and BAM - filler content delivered! Ever notice how corporate blogs are typically home to the most useless "content" on the web?
The original article is obviously written for branded email marketers... Not 3rd party list managers.
Except for this statement: 1. What are my acquisition sources? Prioritize your acquisition sources based on risk and data quality. Sources such as new point-of-sale and sweepstakes should be considered higher risk for deliverability issues and spam complaints. It's like I said...a poorly written article that offers no useful information and isn't even clearly defining its intended audience.
Branded marketers still have different sources / channels even if they are not 3rd party. The article is useful just as a laundry list. It is up to the reader to decide how to apply it, but categorizing your data is pretty basic stuff.
Well indeed this article isn't geared towards 3rd party, nor is it geared towards a seasoned mailer. Thanks for ripping the sh*t out of it DG
You are welcome! I rarely hesitate to call BS when I see it, but trying to find fault everywhere is just facepalm-worthy.
It's not useful since it's not telling you anything important and it is corporate filler content. Categorizing your data is generally a recipe for failure because, as I said before, all people are prone to buying/signing up for things at different times for reasons that do not always make sense. Fat people buy diet products but they also like coupons to McDonald's. It may seem "logical" to only send that sexy lingerie offer to a list of women, but guys may also be interested in buying it for their gf/wife/cockholster. Sending to the list as a whole maximizes your potential. The only real division you'd want in your lists is clickers vs non-clickers...the rest would be identifying what offers respond better to your list AS A WHOLE based on the particular vertical the offer is aimed at.
If you know how to do it you can feed any inputs into a data mining algorithm and find some patterns of predictive value.
Dold, you are a dick but I respect it and kind of look forward to your posts. LOL. Anyways, the craziest stuff works. I often refer back to this story... A few years ago when I was working for a big email co., a friend of mine called me to ask to test an offer. The offer was for some pill to stop peoples ears from ringing. Within a week or two he had a couple thousand sales. Who would have thought that many people had problems with their ears ringing...?
Airin appreciates respectable dicks and she gets what I'm saying here. Even the best predictive algorithms would overlook something like that, which is why you don't want to chop your lists up - leave categorizations to the offers you send.
I've tried both scenarios multiple times...I've tracked all my clicks and opens in general and in the individual categories/offers that they are opening and/or clicking in. Segmenting out and then sending to those specific "targeted" lists always underperformed sending to the entire clicker file for the reasons Dold states. Unless you are paying a CPM for sending thru an ESP, blasting out the entire list will generate the actions from the "diet coke" clickers and additional ones from the "McDonalds" clickers that would have thereby been omitted. It reminds me of a phrase a wiseman once told me.."If you're not mailing the data, someone else out there surely is" Noone's list is unique. Blast away.
We tried creating a test list from a random 10 - 25k record sample of each of the lists we managed -- pulling 1 record out of every 10, 2 records and then skip 10 and pull another 2, etc. We tried mailing singly to those segments, putting different segments together as a master test list, etc. Whatever we sent always under performed (based on eCPM) than to sending to the first/ last/ middle consecutive 100k records. After a couple months of driving our database admins crazy, we finally gave up and tested the old fashion way...throwing shit on the wall and seeing what sticks. While we had a lot more capacity than a lot of mailers, we tested at least 100 new offers every single month. We were lucky if 1 in 10 hit our eCPM minimum. When we found the offers that hit, we put them on a rotation through all of our lists. The lists that the offer performed on we would hit with that offer every 5 - 7 days until they stopped performing.
Wise words indeed. Yes, I agree people can go a little crazy with "data segmentation" best to keep it loose. Segment the openers and clickers but I would never limit an offer to specific list UNLESS (this only applies to branded mailers) if the recipient only agreed to receive offers from party 1 and not party 2, since complaints will likely increase. However, what I suggest is adding addition "promo" offers in party1 email newsletter, like they do on websites i.e if you like this offer you may like offer 2,3 or 4 ....ect, thats just one example of how you can introduce new offers to a variety of lists.
