Lookalike functionality is needed for the clients to find accounts that share similar characteristics with a list of accounts defined by them. Such account lists could be for example client’s won sales opportunities or companies with open opportunities in their CRM. The need for the client with lookalike is to easily create a larger list of accounts and to identify accounts that they were previously unaware of. These accounts would be used for targeting in advertising or for sales outreach.
Practically lookalike is an easy way for the customer to easily create a big enough account list for ABM advertising and also show internally that N.Rich platform provides value (finding previously unknown accounts).
In its most basic form, the lookalike algorithm is based on a positive list of accounts (domains) and it’s able to identify similar accounts based on firmographic parameters (company size, industry, and country).
Technographics may also be potentially useful, i.e. technologies used by the company, but with those, there are several issues, such as many technologies being very common (e.g. Wordpress as CMS) or completely irrelevant to client (e.g. if client sells ship engines, what’s the role of CMS in determining account potential)? We need to experiment with this area to see if it’s actually providing value.
We should identify accounts in the dairy industry as both accounts have this industry. The best matches would be dairy companies with 10 000+ employees in Finland and Sweden. Next category would be 10 000+ employee Dairy companies near to those countries, e.g. Denmark, Norway, Russia. And after this a bit smaller dairy companies.
There is low similarity between these accounts in terms of firmographics, but we should be able to identify that both companies are using demandbase as a technology. So our algorithm should find other companies that are using demandbase as well.
UI flow has three stages:
Define list source (either domains or existing segment)
Define how many companies you want to have in target list (this will define how accurate our algorithm should be)
Show client the resulting parameters and let client decide which ones are relevant, i.e. uncheck the box (e.g. if we find out that Wordpress as CMS is a factor, and it’s not relevant for the client he could disable this parameter).