In current occasions, we use navigation apps like Google Maps and MapMyIndia to discover new areas. From a shopper’s perspective, they’re certainly good and supply all related info on site visitors, eating places, clinics, and so on at a given location. However is there a Google Maps-equivalent for a businessman eager to arrange a retailer in a brand new location? Who does the recce for him? Who brings in intel?
GeoIQ does that and way more. A locational-intelligence platform primarily based in Bengaluru, GeoIQ has achieved 10x development in its ARR over the previous 4 quarters and is on its option to registering extra 5x development in ARR within the subsequent two quarters.
In an unique interplay with Analytics India Journal, CEO Devashish Fuloria speaks about GeoIQ’s enterprise fashions, flagship merchandise, the challenges that the agency faces and likewise the longer term plans.
AIM: What’s the story behind GeoIQ?
Devashish: Round 2018, we have been discussing quite a lot of issues round knowledge science. One of many largest retailers in India was making an attempt to implement their plan of opening a series of 5,000 grocery shops throughout India, however confronted issues in figuring out essentially the most appropriate areas for his or her shops past metro cities. Usually, entities depend upon info coming floor up in such instances. Nonetheless, it turns into tough to know which info is sweet sufficient as a result of entities usually have no idea the whereabouts of a location. Now, while you arrange the shop, it’s an enormous expense. So, if it doesn’t work after one 12 months, you’re dropping quite a lot of enterprise worth and capex.
Therefore, we got here up with the thought of offering location-related info in a centralised method. We began testing the information for numerous areas. We checked out authorities knowledge, public information and the web and finally layered all the data on the map. As we speak we inform companies every part about an deal with on a avenue with out really being there – whether or not a location is reachable, whether it is dangerous, would there be demand for an costly attire, and so on. We offer solutions to such high-value queries in easy numbers. When a shopper enquires a few location, we give them a rating between 0-100, which signifies how good that location could be for the enterprise.
AIM: That’s attention-grabbing! Please elaborate on a number of the newest use instances supplied to your clients, notably in deploying AI/ML?
Devashish: Our use instances are throughout industries, the biggest being in fintech, retail and e-commerce. The largest use case for fintech is danger prediction. Earlier than disbursing loans, fintechs must know concerning the credit-worthiness of their clients. When making use of for loans, clients present info like PAN card particulars, bank card numbers and so on. Based mostly on location-specific info derived from these particulars, we predict how credible a possible buyer could be.
For e-commerce, we remedy the priority associated to return-to-origin by augmenting their person knowledge with hyperlocal intelligence. Different use instances embody correct affluence prediction, fraud prediction, claims propensity, collections mannequin, and lead prioritisation.
AIM: What are the challenges you’ve confronted thus far and the way is GeoIQ addressing them?
Devashish: After we began out, we realised there was a powerful want, but a restricted understanding of this knowledge. We needed to educate our customers, unsuccessfully at occasions, on how greatest to make use of this huge knowledge repository. Enterprise analysts all the time regarded for particular numbers that they thought have been necessary. This was finally addressed. Now we offer purchasers with scores.
Then there have been challenges like sourcing precious, correct, and high quality real-world knowledge. For instance, good-quality knowledge for Indian boundaries weren’t simply obtainable, nor accessible. Additionally, the pin code, metropolis, state, and different boundaries weren’t very clear past administrative limits. Subsequently, we constructed a GeoAllocation engine to outline the boundaries higher (20% extra correct than present boundary definitions) and mapped the addresses on these new boundaries.
One other problem was with respect to the sanctity of knowledge. We needed to test and validate knowledge factors from numerous sources to make sure the accuracy and truthfulness of knowledge that we had gathered from public knowledge sources.
At current, knowledge discovery is a giant problem. Figuring out which knowledge is beneficial and impacts a use case instantly is a herculean activity and sometimes lands on trial and error. The NoCode ML platform has been created to resolve this downside. The answer has accomplished the beta part and is quickly to be launched. Companies would be capable of create a number of fashions very quickly and experiment with knowledge attributes to establish the very best match for his or her use case.
AIM: So, in a method, GeoIQ is making an attempt to handle the issue of unhealthy knowledge with the NoCode ML platform. Unhealthy knowledge is a worldwide downside. How else do you propose to sort out the menace?
Devashish: Knowledge is sweet. We don’t imagine in unhealthy knowledge. Unhealthy knowledge is sweet knowledge not offered in a structured and consumable format. Sure, we don’t deny that there are lots of issues with knowledge while you have a look at firms’ databases. However you possibly can’t remedy a number of issues. Thus, we’re specializing in one particular space, i.e. location, that helps us bind quite a lot of completely different datasets collectively in a single cohesive method.
AIM: What know-how does GeoIQ use to analyse a broad knowledge stack to offer customised options to its purchasers?
Devashish: We use our proprietary algorithms to rework real-world knowledge from 600+ sources in a structured format, categorised underneath 3000+ attributes. Now we have constructed superior machine studying capabilities that assist us present customized fashions to the purchasers, primarily based on their use instances. The explorer and APIs, our customary choices, are backed by ML capabilities. Explorer helps a person to get details about a selected location or examine as much as three areas at a time for a selected attribute. APIs may very well be instantly embedded within the code to enhance it for location knowledge.
AIM: GeoIQ raised USD 2.25mn in funding just lately. How do you propose to channel these funds?
Devashish: As a product-led firm, our core methods are round product growth. We’re channelling part of these funds in direction of constructing superior tech capabilities and a group of expert human assets.
Furthermore, we wish to take our methodology and construct an identical system for various geography. As of now, it’s the US. We’re opening our platform to knowledge scientists, the place they’ll create their very own location fashions and get particular location solutions associated to the US.
AIM: Why US, and never the bigger area round India?
Devashish: Southeast Asia has very related issues to India, which is principally that we’re all knowledge poor. Southeast Asia, for us, possibly twenty geographic entities, which suggests it’s a must to discover 20×500 sources of knowledge to construct up a base. However US being one huge geographic unit, you possibly can scale up the information in a short time, and likewise, the information is definitely obtainable.
AIM: At current, GeoIQ provides three merchandise – No code ML, API and Explorer. Are there plans to develop the product vary? What’s within the pipeline?
Devashish: Greater than increasing into new merchandise, we are attempting to develop our options for brand new use instances, supply extra newest knowledge, and enhance ML fashions. Each new knowledge supply that we add enriches our choices a step additional. Our options are seeing large adoption. New firms with very particular and area of interest downside statements are reaching out to us. We’re enthusiastic about these and are increasing our use case eventualities.
AIM: In India, small companies usually undergo because of location points. From the CSR perspective, do you’ve plans to have interaction with that part of society?
Devashish: We have already got a freemium mannequin for small companies to entry our options with a sure variety of free credit. Additionally, some primary details about cities and villages and streets is being supplied without cost.
AIM: Do you suppose discussions round knowledge regulation would impression GeoIQ’s operations?
Devashish: I feel it was a really aware alternative we made 4 years in the past that we’re not going to get into private knowledge. It’s anonymised info that we take care of. Authorities knowledge doesn’t identify individuals. Info like Nielsen’s knowledge talks about spent patterns at a location and doesn’t have any private info. We don’t know who Particular person A is. That info doesn’t exist within the system. If something, GeoIQ’s methodology units a template for a way knowledge needs to be used throughout a number of techniques whereas remaining ‘privateness first’. Subsequently, there can be no hostile impression of the altering knowledge rules on GeoIQ’s operations.
AIM: What’s the roadmap forward for GeoIQ? What growth can we count on when it comes to new options and companies?
Devashish: Fashionable AI techniques that exist inside firms are primarily based on what customers are doing with their apps. No one is aware of what the customers are doing in the true world. For instance, Particular person A is simply an entity for Amazon, and all personalisation will occur primarily based on A’s motion and behavior on the app. It’s seemingly that A’s neighbour has related preferences as A, however Amazon doesn’t find out about them. We’re searching for to faucet these kinds of frequent interactions and make them a part of the trendy AI techniques.
From a roadmap perspective, we’re first concentrating on high-volume transactions, thus participating lots with fintechs, insurance coverage, and e-commerce gamers. We’re progressively growing our verticals in India. When it comes to new options or companies, we are attempting to strengthen our foothold within the insurtech and e-commerce sectors by addressing extra use-cases and at a bigger scale.