Powerful community detection
We’ve spent 15 years working with leading data scientists to create the first tool in the world to track offline communities. Why?
- Offline communities are the number one influence on consumer buying behaviour
- Grow strategically by harnessing community influence in the real world
- Maximise your marketing efforts by growing community value
Pandemic modelling, turned upside down
Our data scientists took pandemic science and flipped it on its head. The Herdify team use that type of modelling to predict where word-of-mouth is taking place in the real world.
Pandemic modelling is about predicting where a virus will transmit between people. The same models can be used to predict how brand messages will transmit between people.
An AI engine, fuelled by GDPR-compliant data
Using this science, Herdify can predict where people are talking about your brand in the real world.
We don’t rely on social media data. Instead, we take your sales data and look for patterns to see the location and frequency of how these messages are being transmitted.
Let’s dig a little deeper
Today there are no options to identify offline word-of-mouth; 60% of conversation happens offline so social listening misses this. Digital retailers have random sales that cover entire countries, so when you plot their sales locations it typically just shows the locations that are highly populated, not where you have community hotspots.
We normalise for population and then baseline the time between sales. We then analyse at the lowest level geography to see if there is a relationship, and then increase the distance to see how tight the community is. Finally, we compare all adjacent communities to see if a weak narrow one is actually part of a larger one.
You could but this is selling yourself short. Firstly, one good customer does not make a community, and perhaps 10 average ones are better. Secondly, a few good customers don’t tell you how wide the community is.
We need customer ID, billing postcode, date of purchase and value (optional).
Demographic data shows the attributes of individuals and the communities they live in, NOT the communities where your brand is being talked about.
Let’s take an example that you sell a product to 30–40-year-old professionals who read the news.
Demographic data can find these people for you, but it will be the same for your competition, meaning you are in a head-to-head an expensive fight.
By finding your communities, you’re finding the spots where people are leaning towards your brand and just need a gentle nudge to purchase.
Profiling existing customers’ locations isn’t new, but community isn’t just about volumes in a location. Communities can be high in influence, but low in volume of sales as much as they can be high volume and high influence. If you just focus on volume, you end up with a high correlation to densely populated spaces.
This is where the time component comes in. There isn’t an “average time between sales” that is a silver bullet, but more an ever-increasing speed. If you think about how a pandemic spreads – 2 people, 4 people, 8 people, 16 people (you get it) – it has a quickening effect. With product sales, the pattern isn’t as obvious, but it’s a similar trend.
