I spent a lot of time explaining the concepts and benefits of Percollate to people at yesterday’s Technology for Marketing & Advertising exhibition in London. The prevailing methodology to social media engagement seemed to be to broadcast to as large a list as possible and then measure the results. I believe there’s a better approach.

Creating a personal relationship with a customer is always going to provide a better RoI than a blanket mailing campaign. The problem is: who do you talk to? There are quite a few people on Twitter these days and you simply don’t have enough resource to engage them all.

Percollate is not an automation tool. You still have to talk to your potential customers. What it does, however, is point you towards a smaller number of potential customers that you may get better results from.

By way of an example, I analysed 250 people who tweeted yesterday’s hashtag (#tfm2012) from within a 5 mile radius of the event. My goal is to find a handful of individuals that I can talk to about Percollate.

Localised Percollate Search for #tfm2012

I filtered the list to those with Influence and Outreach (as measured by Kred) in 2 communities: Social Media and Marketing. This resulted in 117 users – still too many for me to talk to.

Communities Filter

I then filtered based on users’ global influence and outreach to find 4 users – much more manageable.

Top 4 Influencers in Social Media and Marketing

This isn’t the end of the story though. Percollate does not excused me from manual work altogether. What it does is significantly reduce it. Based on their profiles Chris Windley is probably not a potential customer. Whereas, Emma Jane Gray, Kursha Woodgate and Robert Pickstone might be. I’ll be following up with these 3 individuals to see if Percollate might fit well in their businesses.

If you’d like to do your own analysis or discovery why not take advantage of our 2 week free trial?

Last week O2 had a bad day. You can read why here, but suffice it to say their user-base was very angry and very vocal. O2’s Twitter stream was bombarded with questions from customers who wanted answers – and quickly.

Looking past the initial infringement, it seems that O2 tried to do the right thing. They were listening to their social media streams and doing their best to engage and comfort users. Unfortunately, as Tom McCann wrote in his blog ‘Tweet limits pose a social CRM issue problem for O2‘ they simply couldn’t cope with the volume.

Faced with the problem of not being able to talk to every disgruntled customer what should they have done? Hindsight being what it is, they should have turned to Percollate. Within minutes they could have identified a manageable number of users with either strong influence or large followers. Spending valuable resources on these users would have provided a higher ROI than just responding to people in chronological order. They could have converted these pivotal users into advocates and asked them to help with quelling the masses.

If you’d like to know more about how Percollate can help in your marketing or PR organisation please visit our website for a free trial.

Percollate now lets you sort, filter and group Twitter data based on Kred influence metrics.

At Percollate we’re very excited by the growth in companies offering measurable insights into social media. Since ‘Day 1’ we’ve included this valuable data in our application, starting with Klout. Soon afterwards we added PeerIndex and today we added Kred.

We love all these companies but our goal is not to tell you who’s “best”. From experience we know that all our users are individuals with their own understanding of the merits of each metric.

So, what do you get?

We’ve initially added 4 Kred metrics into Percollate.

Kred Influence

Kred Influence is the measure of what others do because of you. Influence increases when others take action because of your content. Your Influence score increases when someone retweets, @replies or follows you.

Kred Outreach

Kred Outreach is the measure of your generosity. Outreach increases when you retweet, @reply, or follow a new person. As you accumulate Outreach Points, you move to a higher Outreach Level. Because Outreach Points are a reward for being active and benevolent, your Outreach Level never goes down.

Top 5 Kred Influence and Outreach Communities

Here we list the top 5 communities that users are either influential in or have outreach in.

We’ll cover some usage examples in our next post. In the meantime, go and take a look for yourself with our free trial.

We’ve just deployed an update to percollate that will allow you do to full-text searches. Once you’ve loaded a collection of users you can search much of their meta data and immediately filter on the matches. The searchable categories are:

  • Name
  • ScreenName
  • Location
  • Latest Tweet
  • Description
  • Klout Topics
  • PeerIndex Topics

This is a result of searching twitter for the keyword ‘marketing’…


You’ll now see a new text box in the top left hand corner. Simply type and hit enter or click the search icon…


Searching for the word coffee reveals just 3 matching users.


Why did these users match though? Alessandra has coffee as one of her PeerIndex Topics…

alebrandcare (Alessandra Colucci)

Alivia mentioned coffee in her most recent tweet…

AliviaYoungwing (Alivia Young)

And Claire cites coffee in her profile description…

Clare_TheBian (Clare)

Earlier today we’ve published an update to percollate. You can now search for information within a geographic cordon. This enables you to get some filtering done early, before the data has even hit our platform. It’s also very flexible and easy to use.

Firstly, this new feature is only available on keyword searches. You’ll see a new map pin appear on the right hand side when you change the tabs to ‘Keyword’…

Keyword Search showing Map Pin

Clicking this pin will open the location filter…

Location Filter

The text box will allow you to search in a very free-format way and display the matching locations in the list on the left. For example, city…


This gave us all the Norwich’s in the US, but I want the one in the UK. So, I can just make my search a little more specific…

Norwich UK

I could also search for a postcode, zip code or full street address…


One added bit of ‘cool’ is the ability to search for known places of interest…

Silverstone Circuit

I used that previous search for the ‘British Sports Car Manufacturers tapping Formula 1. A Marketing Use Case.‘ post.

To conclude this post I’ll re-enter ‘norwich uk’, select the one item in the list and then click ‘Accept’. This location filter is now locked in and will stay that way until you ‘Clear’ it. Finally, searching for the hashtag #socialmedia with this filter gives the following 15 results.

SocialMedia within 50 Miles of Norwich UK

The inclusion of this one feature has added so much power and flexibility to what marketers can do with percollate. If you’ve not sign up for our free early access program it’s not too late.

In this fictional use case I will play the role of a marketing agency for a UK-based sports car manufacturer such as Caterham, Westfield or Lotus. I’ve been tasked with finding a handful of carefully selected people to invite to one of my client’s many track days. For the best ROI I want people that are either potential customers or may be able to influence others to become customers.

During today’s Formula 1 race in Abu Dhabi I put our fledgling location-based search through its paces. My starting point was to look for anyone mentioning the #f1 hashtag within 25 miles of the Silverstone Circuit in the UK. Our new ‘More…’ button allows the user to repeatedly get the next page of results. So I quickly found the most recent 300 twitterers that matched.

300 People Mentioning #f1 near Silverstone

Looking at the number of followers it seems I have an outlier. This is an official Mercedes account which I’ll exclude with a simple filter…

Now, even though my search includes a location element, the results contain people who retweeted messages from other people in the correct area. So next I’ll use percollate‘s Location filter to improve my results…

Location Filter

This turns out to be quite an influential group, so filtering them based on a few Klout scores doesn’t make an enormous difference…

So, I’ve now got people from the right location, with the right level of general influence talking about the right topic. What I’d also like is people that are specifically influential in cars and motorsport. For the final step I’ll use a combination of Klout‘s Topics and PeerIndex‘s new Benchmark (Sport) metrics to reduce our sample set to 12…

These people have a better-than-average chance of being ‘petrol heads‘. They love motorsport and, sometimes not so, secretly aspire to become a racing driver. The most accessible form of this is through race-focused cars from manufacturers such as Caterham and Westfield.


Using a location-based search of a prominent motorsport event I was able to discover a dozen individuals who enjoyed, and were influential in, sports cars. Brands such as Caterham, Westfield or Lotus could easily invite these folks to one of their many taster track days at Silverstone. If they have a great day then they’re extremely likely to shout about it to their followers. Inviting such a targeted audience should provide a much greater ROI.

If you’re a wine producer, seller, distributor, marketer or consumer you’re probably watching WineFuture Hong Kong now. You might also be tuning in to the stream of tweets containing the hashtag #wfhk11. Here’s an example of how you can use percollate to discover relevant people within that stream.

What we did was grab the unique twitter users that were responsible for the 100 most recent tweets containing #wfhk11.

From here we can see the most influential European tweeters as measured by Klout:

We can also filter those that both Klout and PeerIndex find influential (reducing the number by 2):

But that’s not the end. Here are the 3 people who, according to PeerIndex, are specifically influential in French Wines:

And here’s who you should talk to about Wine Tasting:

We’re still taking requests for early access to our prototype. Signup today and you could be doing your own analysis on #wfhk11.

Percollate Demo: targeting competitors. from percollate on Vimeo.

This is the first in a series of posts where we’ll show how we’re dogfooding @percollate. We’re trying to spread the word, create some buzz and talk to potential customers about our new product; but where do we start? Our tool helps marketing professionals discover great users to start a campaign with. So, that’s what we’ll do ourselves.

@thirstforwine mentioned that the #dellb2b conference was spawning quite a lot of chatter about Klout and PeerIndex. So, let’s run a query to find everyone who’s mentioned #dellb2b recently…

This gives us a collection of just 40 Twitter Users…

We can graph these users to show their distribution based on, say, Klout Score

Next, let’s discover who is influential in certain topics. For this we’ll filter on PeerIndex Topics ‘marketing’ and ‘social media marketing’ and then sort the users by their PeerIndex score

If we want to understand more detail about these 11 users we can simply zoom in…

Because percollate doesn’t believe in ring-fencing your data you can export it  to standard csv format.

And that’s it. We’ve now identified 11 specific users involved with the #dellb2b conference that we’ll be contacting and offering early access to our prototype.

Finally, if you’re a social media marketing professional we’d love to talk with you. Get in touch via our website and we’ll set you up with early access to our prototype.

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 3 other followers

Twitter Updates