Using robots and algorithms to build a campaign planner and issue salience app
Something a bit different this week!!
Having worked in media and comms for about a billion years, there are two questions I regularly get asked by people:
Contacts: Which journalists are best to talk to get “x” story in the media? and
Planning: How do you plan proactive campaigns work work, given it’s so hard to know what issues are going to drive the media and political agenda tomorrow?
I’m always happy to share contacts, but in a world of “key performance indicators” and “impact” I was intrigued to see if there was a more objective way of assessing who to work with in the media.
So a year ago I set myself the task of trying to answer these questions using tech. I learned a bit of coding and did the following:
Built robots to harvest news stories by influential journalists (political, economics and business journos mostly)
Used natural language processing algorithms to analyse media content and assess “issue salience”.
Knowing what issues journalists are interested in is obviously helpful for targeted outreach. But more than this, modern techniques like natural language processing allow us to:
know what issues are salient among target journalists (eg “inflation”, “cost of living” etc)
understand who journalists are talking to in order to make their stories stand up (sourcing strategies), and
draw conclusions about the issues, actors and institutions that are dominating the media debate and - more importantly for progressives - whose voices are missing.
Last year I ran an experiment tracking the work of a set of economic journalists to see what I could glean about the above. This briefing note explains what I did and some of my findings. Below is a word cloud of some of the issues that came out of the exercise.
Over the last few weeks I’ve been monitoring and analysing coverage of the Conservative Party leadership contest. I selected a number of key political journalists to track (from the BBC, Guardian, The Sun, The Mirror, Daily Mail).
I’ll use my next newsletter to update you on what I found!
Using tech for campaign planning
There are loads of media and campaign planning services out there. But in my experience they can be expensive and rarely well tailored to the work I do. Part of the difficulty is that the planners that do exist are often consumer focussed. There’s probably little margin in providing a campaign planning service for NGOs, think tanks or niche campaign groups.
Yet, there is a wealth of planning data out there. In my experience the news agenda across politics, economics and social policy is heavily influenced by set piece or planned moments (from government reports and statistical releases to parliamentary debates, campaign events, legal rulings, budgets etc). And with a bit of creative thought (and nerdery) the internet can be a treasure trove of data just waiting to be harvested into a campaigns, media and public affairs planning tool.
Below is a screen shot of the planner I’ve developed, it provides a glimpse of the issue salient moments taking place at the beginning of September 2022.
My robots can harvest information from a limitless number of places covering a limitless number of issues (from immigration to health, to the economy). The planner includes a top line that is then categorised by issue, date, details, organisation and a url link so people using the plan can get more information.
Here’s another version of the planner, but this time in a calendar format. It’s just a snapshot (the planner goes out as far as January 2023 though the quality of information out there becomes sparser the further ahead you look).
There are probably a number of ways of presenting the information in a way that’s helpful for campaign, media or public affairs work.
What do you think, is this of use to you?
I’m thinking of developing the tool further and would love to hear ideas from people about the sorts of issues and organisations you need to track in your work. Are planners like this useful? Do you use a planner of your own? Perhaps there are particular issues or organisations you need to track? Let me know and I’ll see if I can return any useful data to be added to the planner (I love a task).
In the next edition of this newsletter I’ll update you on the issues that were most salient in the Conservative leadership race.
*Another example of the use of these techniques was the GameStop Stock Frenzy, this year, when a bunch small scale investors used the online message board, Reddit, to co-ordinate a share buying strategy that annoyed a bunch of hedge funds. It’s rumoured that smart City traders used scrapers and natural language processing techniques to keep tabs on the online noise so they could keep ahead of the mayhem and make a buck. Scrapers are also useful for people who make a living from arbitrage trading).