What is data-driven content?
First of all, let me quickly clarify what I mean when I say "data-driven content"...
I'm NOT talking about creating content based on audience insight or understanding previous successes -- that's a topic for a whole different post, and here at Code we have a Strategy team dedicated to finding that insight and applying it for brands in the digital space.
What I AM talking about is content where data itself is the subject, i.e. content that tells a story through statistics or data.
This could be based on a survey/poll, like this research showing that even Conservative supporters don't think David Cameron is in touch with normal people or this poll showing that James May is more popular than 50 Shades of Grey; it could be content based on data in the news, like this graphic showing that 9 out of 10 big pharma companies spend more on marketing than R&D.
Or it could be something completely different - the possibilities are infinite.
Why bother with data-driven content?
Whether you're working on a big B2B campaign or creating content for a hip millennial consumer brand, there are always ways you can use data to create compelling content that resonates with your audience and gets you featured on the sites that matter.
Data can give your content or PR the hook it was missing, and massively improve your outreach results.
Plug the phrase "survey finds" into BuzzSumo, and you can see the kinds of sites that publish these stories, as well as the impressive social engagement. (You might also infer that religion, sex and cats are all good topics to incorporate in your content plan...).
Data can be a powerful way of telling an important story; it can help you become a true authority within your particular niche. But it doesn't have to be serious. Back something up with a survey or some hard data, and with a bit of imagination, you can make almost anything newsworthy. It's all about finding a story. If you want a repeatable tactic for securing coverage and links from big news sites, then data is your friend. Of course, though, with great power comes great responsibility, so you have to be sure to use it wisely!
So how do you get started with data-driven content? Where do you get the data? How do you present it?
In this guide, I've collected all the resources you'll need to start making data work for you. You don't have to be an Excel ninja or a maths genius -- how you work with data can be as simple or advanced as you feel comfortable with. The most important thing is understanding how data can help you tell a story, and communicating that story in the clearest, most compelling way you can.
Sourcing data: tools, techniques and resources
In this section, I've listed the best tools, techniques and resources for sourcing data -- this is the raw material you'll be working with, and there are plenty of options to explore.
Surveys & polls
Here are some of the best tools for gathering survey data.
Toluna QuickSurveys (https://www.quicksurveys.com/)
Toluna lets you build surveys easily -- you can survey your own database for free, or use their panel of six million consumers worldwide on a pay-as-you-go basis. As an example, 1,000 responses to one question with no demographic targeting will cost £200, whereas 2,000 responses to five questions targeting UK females aged under 30 will cost £3,500 (they have a pricing calculator here).
It's also really fast. If you need lots of low-cost, high-quality responses to a survey / poll in less than 24 hours, this is my favourite option by far.
Google Consumer Surveys (http://www.google.com/insights/consumersurveys/home)
Google seem to have a tool for everything, and their consumer survey platform is well worth a look. You can get up and running in minutes, and generate responses for as little as 10 cents each.
Segmentation is based on age, gender and location, and formats include multiple choice and open answer. You can also ask screening questions to narrow down the respondents you want.
Want to see an example of a story based on Google Consumer Survey data? How about this recent piece in the Huffington Post, based on the finding that 27% of workers admit to binge drinking on business trips.
If you need to survey a more specific demographic group (e.g. 'car drivers' or 'IT decision makers') SurveyGoo is a good option. Like the other two options above, this is a DIY, self-serve platform that makes it easy to design and launch surveys online.
You can check how many panel members are available that meet your criteria and what the cost for your survey would be here.
Google puts a load of free tools at your fingertips. Here are my favourites...
Google Trends (https://www.google.co.uk/trends/)
Google Trends is one of my favourite tools -- you can use it to visualise changing search interest over time for topics or keywords you're interested in, and compare up to five things against each other.
In the example below, you can see how searches for 'Clarkson' have rocketed since his recent fracas, or 'steakgate' as it has inevitably been christened.
Google Correlate (http://www.google.com/trends/correlate)
This is a less well-known tool from Google that lets you find similar patterns in search interest for disparate topics, as well as finding search patterns that match trends in your own data. The uses for this aren't quite as obvious, but it's an interesting tool and well worth having a play.
Google Ngram Viewer (https://books.google.com/ngrams)
Google's vast digital library of books dating back hundreds of years is a fascinating data resource, which they have made easily searchable using the Ngram Viewer. This tool lets you see how frequently specific phrases have appeared in literature over a particular period of time, and allows you to compare multiple phrases together.
In the example below, you can see how mentions of 'wine', 'beer' and 'whiskey' in literature have fluctuated over the last 200 years or so.
Google Public Data (http://www.google.com/publicdata/directory)
This is Google's searchable database of public data, with datasets from bodies like the World Bank, OECD which you can segment and visualise using Google's user-friendly interface -- you can create line graphs, bar charts, map visualisations and bubble charts.
Here's an example I created comparing % of GDP spent on education for two world regions:
Google Analytics (http://www.google.com/analytics/)
If you're a digital agency like us, you probably have access to a wealth of data across your clients' Analytics accounts. Aggregate that data, and you can start doing some really interesting analysis and identify key trends, like we did in the [stats piece] I mentioned earlier.
But even at the individual site level, there may well be some interesting data that you can pull out about your business and the people that interact with you online which could form the basis of some content / PR. Pornhub are the obvious example of a company doing this exceptionally well -- their insights blog makes undeniably fascinating reading, and has generated masses of coverage for their brand.
UK Government & public sector
Thanks to Freedom of Information laws and the Open Data Charter, a treasure trove of government and public sector data is accessible to all. You can add the value by analysing and interpreting the data, or combining it with other data for added context.
GOV.UK Statistics (https://www.gov.uk/government/statistics)
Search government statistics based on keyword, topic, or department. You can also see announcements for recent and upcoming releases, which is great if you like planning ahead.
On the government's open data site, you can browse datasets based on theme (e.g. environment / health / education) or search based on keywords. It's not perfect by any means, but it's definitely worth knowing about.
Freedom of Information Requests (https://www.whatdotheyknow.com/)
Can't find the data you want but pretty sure it exists? You can request it directly from the relevant government department or public sector body via a Freedom of Information Request. The WhatDoTheyKnow site makes this ridiculously easy -- just make sure you're crystal clear about what you're asking for, and keep your requests simple and to the point.
Legally, you should get a response within 20 working days. Think like an investigative journalist -- hunt down the information you want and find the story that hasn't been told.
I won't go into detail on each of these, but if you're looking for government/public data outside of the UK, check out the following resources:
Social media is an amazing data resource, but you'll probably need an expensive tool like GNIP or a developer skilled at data mining to make the most of it. This is not really an area I have explored, but there must be so many fascinating insights hidden in the datastreams of Twitter et al that are just begging to be discovered.
If you think you might have the technical know-how to jump into this, then check out the Twitter API documentation and let me know how you get on!
If you haven't already played with the YouGov Profiler, make sure you take a look at this tool. It won't give you numbers, spreadsheets or graphs, but will give you some very nicely presented insight about what differentiates fans of a specific brand, person, or thing from YouGov's comparison set.
The output is an illustrated 'profile'/persona with seven sections. For example, you can see insights based on 6,204 fans of David Cameron, telling you the obvious things like they're likely to be well-off and based in the South of England, as well as some quirkier details, e.g. they like to eat roast pheasant, lobster thermidor and beluga caviar. YouGov do offer a far more advanced, paid version of this tool, but it's not cheap!
Tools for analysing, organising and manipulating data
In some instances, such as head-to-head 'this versus that' polls, the data you get won't need any more work - you know that 70% of people polled voted for 'this', and 30% voted for 'that'. Simple.
But in many cases, your data will need cleaning, formatting, organising, manipulating and interpreting before you find your story and your killer fact. This is where some spreadsheet skills will really come in handy -- VLOOKUPS are your friend. Alternatively, buy some dougnuts for whoever can help you out with this in your team, or outsource your data analysis if you need to.
The tools and techniques you'll need will really depend on the complexity of your dataset(s), but you should be able to do everything you need with the following:
- MS Excel / Google Sheets -- Excel / Google Sheets should usually cover most of your needs
- Google Fusion Tables - Filter, summarise and merge datasets, and visualise them in embeddable graphs, charts and maps -- check out the tutorials here
- Open Refine -- A power tool for cleaning up messy data.
Data visualisation formats & golden rules
Once your data is clean and you have worked out what your story is, it's time to think about how to present it. You may decide that a simple blog post/press release is adequate, but often the best way to communicate your story will be using some sort of visual format.
Last year, I attended a data visualisation masterclass run by Rob Orchard and Christian Tate from the Slow Journalism Company (check out some of their work here). They broke down the potential data visualisation options into five formats, and suggested 10 golden rules.
The five possible formats for your data visualisation are illustrative, proportional, timeline, map and list.
Image live-sketched by Chappel Cartoons.
Illustrative = images to convey small, bitesize facts (like [this graphic] from one of our recent blog posts).
Proportional = charts, graphs, and clever things like this epic infographic showing the number of sharks killed by humans each year or this interactive visualisation of the what the contents of an iPod would look like in vinyl form.
Timeline = data arranged chronologically to tell a story of change over time, like this piece on Robin Williams' acting career.
Map = geographical data overlaid on a map, like this map showing where in Europe belief in God is highest and lowest.
List = basically, tables (like this 'Stroganoff the menu' piece, for example).
An infographic could combine all of the above.
Here are Rob and Christian's 10 golden rules for data visualisation:
- 80% of brainwork is done before you begin
- It's all about context
- Go Goldilocks on scope
- Make sure your data are credible and correct
- Work out what the story is
- Make your approach appropriate to the medium
- If it's not working, go deeper
- Make sure your data are comprehensive within a defined area
- Work out your killer fact
- Package it up nicely
Data visualisation tools
How you go about presenting your data will depend on:
- The medium it will live on
- The type of data
- The story
- The killer fact
- Your resources/skills/budget.
I've divided some of the options for visualising your data into beginner, intermediate and pro.
Piktochart -- create infographics from templates with a simple point and click interface; basic but a good place to start.
Infogram -- another easy interface for creating infographics, as well as an impressive array of charts and maps.
Wordle -- the easiest way to create word clouds -- perfect if your data is text-based, such as written responses to a survey question.
Tableau -- this is a more sophisticated visualisation tool with a wealth of options, but still available as a free package -- check out this guide to creating data visualisations using Tableau's Story Points from Journalism.co.uk.
Google Fusion Tables -- I mentioned this one earlier, as a data formatting / analysis tool, but it also has some nice features for generating graphs, charts and maps from your data.
Adobe Illustrator -- if you're creating something from scratch, this is usually the tool of choice.
SketchUp -- want to move beyond 2D visualisations? Why not take things to the next dimension with this software -- "the easiest way to draw in 3D." Here are some examples of 3D geographic visualisations to give you an idea of what this tool can do.
Google Charts -- another one for the coders... Create interactive tables and charts with complete customisability using Google's libraries, plug them into real time data, make different elements interact with each other, and do lots of other fancy things.
Over to you
Whatever your experience with data-driven content marketing, hopefully you've picked up a few new tips from this guide.
I'm always interested to see what other people are doing in this space, so feel free to tweet me @joelstein about what you're up to.