Too much data?!
The truth of the matter is that it’s usually not easy to get good data. Even after you think you’ve made all the hard decisions, you can’t just wait as the data flows in. There are a number of confusing details to work out and pitfalls to avoid.
To help you deal with the data collection phase of your measurement project, here is your how-to-get-good-data checklist:
Sit down and review your search strings with your supplier(s) at least once a month.
- Make sure that all the terms are current.
- Eliminate any terms that don’t yield any content.
- If you are doing competitive monitoring, make sure that the searches are parallel.
- Be patient. It often takes time and many iterations to get search strings correct.
Have a meeting with all your suppliers to review how impressions, audience numbers and monthly viewers are calculated.
- Eliminate multipliers, and establish consistent and transparent rules.
Accept that you will not get every clip, so don’t try to boil the ocean.
- Focus on what really matters: the media outlets, blogs or other channels that influence your target audiences.
- Get those influentials to an acceptable level of accuracy.
Check your data weekly for the first six months; then once a month after that. Get familiar enough with the results so you know how to spot when something is wrong.
- After the first six months you’ll know on average how often you’re mentioned in The New York Times, the local media, and the key trade outlets. Pick the most frequently covering media outlet and set up a system to check monthly. After the first six months you’ll learn:
- On an average month, the key trade journal mentions your company or the competitors X times.
- On an average month, Y stories about your brand or one of your competitors appears in the most important business journal.
- On an average month, you are mentioned in at least X blog posts.
- At the end of each quarter, look at the numbers of each individual publication. If you are way below or above average on any one of those, dig into the data and figure out why. It may be that the competition fired its PR team and they went invisible for a month or so. But a far more likely scenario is that somewhere a feed is broken and there are bunch of missing items out there.
Do not wait until 48 hours before your results are supposed to be delivered to the CMO in a PowerPoint deck to find errors in the data. Review any qualitative data at least monthly, if not weekly. Both computers and humans make mistakes.