The Value of a Tweet: $174 per Donor

A Mid Campaign Tweet by Tweep analysis of the Value of a Tweet

Value of a Tweet - Debbiediscovers

First Posted on Medium: The Value of a Tweet: $174 per Donor — Medium

I’ve been fundraising to build a school in Laos with Pencils of Promise. If you breakdown the sources of donations that I received, Twitter is was the other social media channel that I used to promote my campaign to the public.

(The rest were either direct asks or from Facebook but my Facebook profile is was limited to friends and my Facebook page is was only followed by friends.)

Day 30: Total Donations

This means that almost 10% of my donations have come from Twitter.

 
Based on the number of tweets my post generated (15 tweets from the website and 15 retweets from Twitter), the value of my Tweet is $40.67. Compared to recent research done by Klear, “How much is a Tweet worth for a Kickstarter Campaign” where they estimated the value of a tweet to be $14.30, the value of a tweet for my campaign is 2.8X higher.

You can Tweet this: deb.bi/AddsValue-Tweet-This

(Note: This is not a pure apples-to-apples comparison; there are slight differences in methodology, Klear takes total amount raised by the Kickstarter campaign divided by number of mentions, and to level the playing field for campaigns that have been running for different lengths of time, they normalised the value of the tweet by dividing it by how long the tweet has been live for. I exclude all other donations (because I know exactly where they came from, be it a direct ask or from my Facebook account), and only include anonymous donations that are most likely to be from Twitter. I also don’t normalise it.)

Special shoutout to Andrew Coppin from Sydney, Australia, who was my 8th donor (lucky 888 Mr Coppin!) and my first Twitter donor.

Who exactly are my Twitter followers?

As of 11 Nov 2015, my 2,942 Twitter followers are:

2x more likely to have a net worth of >$1M

2X more likely to have gone to grad school

2X more likely to have a home value greater than $500,000

However, if you look at average household income, there is only a 5% difference between my followers and the rest of Twitter for the highest income category.

  • My guess is that this is because most of my US followers are likely to be Silicon Valley Startup / Entrepreneur / VC / Investor tweeps (inferred using Wireless Carriers AT&T, interests, lifestyle)
  • This is in line with a bio search using Social Rank and Followerwonk, though I can’t estimate the total number
  • For those type of Tweeps, net worth comes from stocks or carry and not from average household income

If you are a US marketer, the amount of data that you can derive from Twitter audience insights is pretty amazing

 

Compared to the overall Twitter community, my followers are also:

  • 20% more likely to buy premium brands
  • 13% more likely to lead natural living + fresh and healthy lifestyles (I bet they’re from California!)
  • 8% more likely to be Kosher (New Yorkers!)
  • 6% more likely to be vegetarian and 4% more likely to be dairy free (OK that’s smells like LA)
  • Most of them don’t do mundane things like buy eggs or milk and it appears that they’re less likely to buy soap either (does shower-gel count? Who uses soap??)
  • But I’m very proud to report that my followers are likely to have good teeth, as they have have excellent oral hygiene habits (17% more likely to buy oral care). You can tweet this: http://deb.bi/FollowerInsights-Tweet-This

But how accurate is Twitter Audience Insights?

1. Country data from Twitter is not accurate: it does not tie in with Regional Data or Wireless Carrier data

  • Many people do not report their location
  • I’ve noticed on a number of occasions that the default Twitter settings for location and time zone are wrong (yes I have 11 Twitter accounts, 8 of which I actively use for experiments, clearer social listening… e.g. @littlecodersSG is the one I use to track what’s happening with Kids Coding around the world; @giftedSGP is one I use to listen to gtchats or edtech chats)

I’ve also noticed substantial discrepancies between what Twitter, SocialBro and Klear report. For USA followers:

  • Twitter reports 40%
  • Social Bro reports 30%
  • Klear reports 56%

2. My initial hypothesis was that looking at Twitter’s Audience Insights on Wireless Carriers should give you a more accurate picture: Based on that (see Chart below):

  • 63% of my Twitter followers use US cell phones (versus the 40% reported in the first column, Country)
  • 19% use UK mobile phones versus the 10% reported in Country
  • 12% use SG mobile phones versus the 7% reported in Country
  • 5% use Airtel which is India

(BTW no one uses the term “Cell Phone” in UK or Singapore, not sure about Australia though, anyone care to comment?)

3. If you add it up, that’s 99% of my followers which does not make sense either

  • Could be that Twitter counts anyone under the T-Mobile Group or the Singtel Group under Wireless Carriers. I asked Twitter but no response — anyone know the answer?

4. Even something as basic as Gender has substantial differences:

  • Twitter reports 75% Male vs 25% Female
  • SocialBro reports 67% Male vs 33% Female (leaving out Unknowns)
  • Klear reports 59% vs 41% Female

These differences are pretty substantial, which leads me to question the effectiveness of geo-targeting (or any kind of targeting) using Twitter Ads, especially for non-US advertisers.

Note: After I posted the update, I’ve had wonderful, responsive feedback from Klear and SocialBro (Now Audiense) — very impressed.

Slightly disappointed that there has been no reaction from Twitter despite multiple reach outs to ask for clarification). And think local carriers like Singtel and Starhub have a lot to learn about social listening.

 

 

 

 

Note: This campaign is over and I completed funding my first school in December 2015.  I then committed to another 2 schools, and have less than $2K left to go so if you found this (dated) post helpful, a $25 donation would make a big difference. You can donate here.

 

Leave a Reply