Blog

There is a lot of confusion out there. Many of the posts and tweets about social media monitoring and brand monitoring these days are asking, “What application should I use? Which app is the best? Who has used what? What do you think?” Very often, the RFPs I receive and the questions I am asked about our platform are pretty surface-skimming. “Does it have Twitter?” is a fine question, but there are so many more tough questions you should be asking all of us.

Don’t even ask me which platform is best. You already know my answer is Scout Labs. I have no credibility there. But, having thought about this and worked on the technology for nearly 3 years, I do think I have credibility in offering up a checklist of things to look for and to think about if you are choosing a social media analysis platform for your company. I will get the list started but I hope that you will add your own thoughts:

1. How many months of live data are you serving in the application? (I need to be able to read all the posts and also see trending graphs).
2. Do you price based on the number of results returned or otherwise limit the number of results that I can see? Or do you return all the search results for my topics?
3. What social media content types do you index?
4. Can I add sources if I don’t see them in your app? How fast does it take to see them in results?
5. How good are your spam filters? What percentage are your algorithms catching as compared to human-scores? How often do you mis-classify something as spam?
6. If you say its spam, can I see what you filtered out? Can I add them back in to my main results?
7. Can I curate the results for my team? E.g. If I think a result is either irrelevant or spam, can I flag it as such and will it be instantly removed from the results?
8. Can I order mentions within a given date range by importance - what’s most important to pay attention to - or is it just by date?
9. What sentiment-related features do you offer? Is it just positive and negative scoring, or can you identify other customer emotions?
10. With regard to sentiment (positive / negative / neutral) how accurate are your sentiment scores? Specifically, how often does your algorithm agree with human scorers? When is the last time you tested your algorithm versus human scores?
11. How are you improving your algorithms over time?
12. Do you do only go-forward sentiment scoring (as posts come in, after you create a search)?
13. Or can you back-fill sentiment scores for old things that were posted before I created my search? If so, is there an addition charge for backfill? How long does it take to back-fill sentiment for all my searches?
14. Does the application support multiple people / departments and groups in the company? Is it easy to add people? How much do you charge additional users or is it free?
15. Is there a way to “bookmark” and tag interesting content as my team finds things and people that we want to remember?
16. Can I annotate things I see with my own commentary and tags, so that my colleagues can see my thoughts?
17. From when a customer posts to when it is live in your application, what is the lag time (by media type)?
18. How long does it take for a dashboard to populate from search creation?
19. Is everything totally automated or do you have your analysts or outsourced labor doing some part of it?
20. How much honing and set-up of the results need to happen to get really good results?
21. How much training time would you say is required to become an expert at using the system?
22. What is the average number of users that are using your platform within a typical company. One? Twenty?
23. How do you price? What’s the cost for an addition person to get access? Is there any limit on the number of search results that you return to me?
24. If you do limit the number of searches that you return to me, (e.g. 10,000) and I do a search on “Google”, say, that clearly has far more than 10,000 results, what do subset do you show me? What am I not seeing?

We don’t mind the softball-inquiries when they are thrown to us. But we know that Scout Labs stands up to very rigorous scrutiny, so bring it on. I hope this makes you a more sophisticated reviewer of platforms and writer of RFPs. I also hope that this forces more data into the discussion about social media platforms and raises the level of discourse throughout the category.