How to Use Data Informed Campaigns

From DevSummit
Revision as of 17:47, 5 May 2015 by Vivian (talk | contribs) (1 revision imported)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Notes

Data Driven Campaigns

In order to be informed, must review old data and determine what is important to use going forward

Social Media monitoring has two pieces: monitoring aggregate data vs individual data

On Aggregate Social Media Data:

  • From Matt/Upwell: The primary tool we use is Radian 6 (social data scraping tool)
    • Surveys the feeds based on your words - sentiment analysis and it doesn't understand sarcasm
    • Issues:
      • It's expensive
      • Sentiment analysis mostly shows neutral (doesn't understand sarcasm)
      • It's designed for brand monitoring
      • Doesn't hook up to a Salesforce contact database
      • For some reason it doesn't follow the attachment in tracking
      • The individual data - Influencer analysis is not great
        • Still building in profiles for those people - doing that manually

Example:

  • Upwell uses data about how Shark Week 2009, 2010 data to create content around 2012 - tell organizations, lead with the message to engage with them the way they know how

We don't always own it - we send them directly to the content that we don't own

Email:

  • Email list - average lifespan of a subscriber is 1 year
  • What works - unfortunately, the things that work are emotionally exhausting like frequency and scare tactics
  • Can test ratio of give to ask ratio on emails as well

Tools for tracking Social Media Sharing:

  • Topsy - share within 30 days (sign posts that pointed at this page) - free with pro version
  • Shared Count (free) - tracks Facebook links - add URL that ends in forward slash - see how many shares
  • Sproutsocial -- $50/month social media monitoring - helps determine timing based on your list and how often they click

Processes:

  • What topic is this on, who is the target, how long will we spend - examine for correlations
  • i.e. it seems like 2 weeks is too long between analyzing ~ run 10-30 campaigns at once in a week
  • Current goal: Volume of conversations online - amplify or extend current things going on online
  • How do you set up AB testing on your influence
  • Figure out the statistically normal level of conversation level and then how you influenced it
  • If you can measure your effect on the spike, you may be able to journalists to give you the story ahead of time

Tell them how to talk about you:

  • In emails, tell people how to share your content
  • Upwell example: we were trying to share a report that had a generic title so everyone who shared it who came up with their own title
  • When we're tracking we find how often it is shared, it is much easier to follow with a distinct phrase
  • It is also much likely to trend
  • This takes copywriting to another level
  • Use a distinct/memorable turn of phrase

Tools we wish we had:

  • Many people use list parsing - with custom drip campaign with each part of campaign.
    • It would be great to have a campaign tracking code for each campaign - workflow tool
  • Tools to see trends across the whole internet:
    • Similar to Google Trends
    • Something like Votizen - see friends have similar political views and whether or not they've voted
    • Looking for developer tools that might integrate
  • Salsa is experimenting with syncing all data to a MySQL database daily to look at changes over time because that format is simpler and more integrate-able
    • Filling in gap of time component
    • i.e. Generic data analysis tools like Creative IO

New campaign trend - tap into positive emotion

  • Thank you campaigns i.e. JC Penney, UPS taking a stand
  • Joycott technique - Carrot - bidding businesses for social change
  • Build community by rewarding positive feedback loop

Measuring interesting data/best practice:

  • Really make sure you're not throwing away data - keep refining the questions you're asking - track people who have left
  • Clicks per open vs what percentage of your list clicks - measure quality
  • How does online influence offline