Hundreds of thousands of people are talking on the web about products similar to the one you plan to develop. Innoradiant provides you with operable knowledge on what they think and how they will react to your innovation.
Think of any generic product or service with a decent diffusion: it is extremely unlikely that you cannot find millions of mentions of it in social networks. These persons are probably using the product on a daily basis and they comment its features, its usability, its value for money etc. They might be enthusiast of certain characteristics while being even unaware of others. Sometimes they implicitly or explicitly suggest changes and innovations. Your next product must benefit all these commentaries.
Comments are millions and are lost into terabytes of irrelevant information. Reading them all is out of question, for time and money reasons, besides being just unfeasible, due to well known limit of human memory. Here is where Innoradiant, with its Voice of the User (VoU) suite, comes into play: understand social media for New Product Development (NPD). We capture user statements about products on forum, we analyse them via powerful Artificial Intelligence algorithms and we return to our customers exactly the kind of information needed to deliver products satisfying underserved user needs.
Selection of the social networks where people speak more about a product.
Automatic identification of features of the product and of the way people talk about them (User Perceived Features)
Semi-automatic connection of user perceived features to standardized product features.
Detection of user attitudes towards known product features.
Operable rendering of user attitudes to perform the right choice for your new product.
For the New Product Development team:
Capture the current key features of the product family, their popularity trend, their user appreciation.
Discover underserved user needs.
Exploit unsolicited user suggestions.
For the marketing department:
Understand how users talk about features of your future product.
Identify influential users and potential ambassadors.
Find the right profiles for most effective marketing actions.
For the R&D department:
Let real user needs drive research and innovation.
Select product features which should be prioritized.
Face user orientation changes which might occur in the development life cycle.
We help our customers to take to the right decisions in the product development life cycle by identifying user attitudes on social networks. Our help is not in terms of consultancy, but is based on the delivery of VoU, a platform which allows product teams to be completely autonomous in the discovery of “killer features” of the new product.
VoU is based on a big data compliant architecture (we do not do much buzz about it, but yes we are dealing with big data!) where several world class Artificial Intelligence libraries have been injected, notably in the domain of Natural Language Processing. Shortly, the main steps to deliver you key indicators about your new product success are the following:
In this phase you select the most pertinent social networks, with the support of Innoradiant experts. These are not necessarily the most popular, but just the most pertinent to the product you are planning. In this phase all relevant spontaneous user interactions are stored locally for being processed.
Product Feature Selection
The starting point of this phase is the identification of what your future product looks like (target product). If it is an enhancement of a previous version, the choice is reasonable easy. If it is a brand new product, something the market never saw, then target products must be identified by analogy. Once the similar product(s) has been identified, thanks to the application of optimized Natural Language Processing algorithms, VoU detects how users characterize it, i.e. which features are more often taken into consideration. Optionally you can normalize these spontaneous expressions to connect the words of users to your internal organization of knowledge.
At this stage you know what are the most popular features of the product under conception/design. But popularity is just one side of the complex polygon of user attitudes. In order to drive your innovation you must be able to answer questions such as “Is this feature connected with an intention to buy?”, “Do they need that feature or they consider it as redundant?” “Do that feature raise enthusiasm or, on the contrary, it is blamed as inappropriate?”… VoU captures these product-related attitudes and disclose them to you as an easy to use dashboard.
Patterns Towards Innovation
At Innoradiant we recognize that sometimes product innovation is so visionary that it is difficult to find “analogous products” to feed the research. In these cases the process become “event-driven” rather than “object-driven”. For instance if you want to deliver a cutter specially fit to cut plexiglas, you might want to track events of people cutting plexiglas, as well as the outcome of that process, rather than features of cutters. We can answer this as well!
Innoradiant “technology stack” is based on a stratification of multiple technologies. The basic idea is to have a background layer integrating big data oriented technologies (basically distributed storage and analysis) with language aware technologies able to add intelligence to the data to be processed (social networks texts). On top of that, the user experience layer facilitate the exploitation of extracted information.
Mixing Artificial and Human Intelligence
In order to be able to extract popular product features and to associate user attitudes to them, the machine must be able to understand the language. This is a typically hard task: years of experience working on analogous tasks prove that it can be tackled only by mixing machine learning algorithms with a human driven, rule based encoding of knowledge.
In recent years there have been spectacular advances in the field of Machine Learning, with the advent of Deep Learning Technologies. Thanks to adoption of a neural network architecture VoU is able to detect and classify human attitudes towards product features. Also we adopt single layer neural networks (aka “Word embedding”) to retrieve terms that should be considered synonyms in specific domains.
As clever as unsupervised AI techniques might be, they have limits in terms of performance and user-driven behavioral change. That’s why our linguists have delivered and refine every day grammars for handling peculiarity of language. This is an important phase of our hybridization technology and a crucial one to achieve high quality. This is also the reason, why serious user voice analysis cannot be language independent. We currently deal with English, French and Italian.
Yes we are Big Data compliant!
For Innoradiant the ability to move into a big data processing framework it is not a fashionable tendency but a need. Even by focusing on forums only, the amount of information to be processed in enormous. Just to have an idea, web places where people discuss about cosmetics are in the order of 14 Million, just for the English, indexed part of the web (often access by robots, such as indexing engines, is forbidden in forums).
We achieve the ability to produce time effective analysis on estimated product success by adopting a distributed secured architecture where we can add nodes in couple of minutes depending on the need of our customers.
We also exploit big data aware technology such as Apache Pig for data mining and Elastic Search for document oriented storage.
We know that your main goal is to understand if the conception of your future product will meet the approval of the market. For this reason our presentation layer let you discover answers to your questions rather than just jiggle with data. Of course you can always go one step further and play around with the complexity of the information we extract, but the point is that you don't have to.
In order to achieve this result we have adopted a user friendly, responsive user interface based on Kibana, which is in turn based on NodeJS and allows real time communication with the data storage (Elastic Search). As for the configuration of the platform, we have rather opted for a more corporate oriented framework such as Java Server Faces.
There are social media which emphasize popularity and self assessment. You know them here are some logos:
However, finding hints on product innovation on those media is like looking for a needle in a haystack. For this reason product related research focusses on forum and thematic online communities (including specialized blogs).
Members of online communities are passionate about the topic
Members of online communities are unbiased: they are not paid for writing reviews: they talk about their activities, and it happens that those in activities they use products/services
Many members of online communities are experts: you should probably listen to their unsolicited suggestions
Online communities are the social network which most closely emulate real life communities
Online communities keep the individual and collective history about an activity; are you sure that for the new version of your product you can ignore what happened one year ago?
Forums contain rich demographic data: The first act of a user participating to a forum is to present her/himself
Forums are most popular than you might think: according to Nativo, a content marketing firm, 20 percent of Americans use forums to discuss and recommend products. Nearly two-thirds of women in online forums make product recommendations on these boards.
Whatever the nature of your product/service, there are hundreds thousands of users writing about it. VoU allows you to select the most appropriate forums in your domain, download the whole history of online conversations and track any daily update.
And of course if you don’t know yet where people talk about your future product, we can help on that!
When planning a new product or an evolution, you must have a clear idea of the perception of the user (or her cognitive model).
You can characterize a product by the property that your R&D department attributes to it. But the only way to successfully model the development process is to look at the product from the same angle as the user.
Thanks to advanced natural language analysis techniques VoU extracts from texts features of products and the way they are mentioned by the user. This is the phase where Artificial Intelligence algorithms play a role in order to provide a meaningful set of user derived features. For instance in the sentence “this wine is sweet” we capture the fact that sweetness is a characterizing feature, but we discard the adjective “available” in “this wine is available.”, as availability is not an intrinsic property of a wine.
The set of adjectives and other kind of phrases characterizing the product are returned to you as the “user-centric cognitive model of the product”. At this point you might keep them as they are or you can semi automatically link them to standard terms used in your organization (for instance “sweet”, “brut”, “dry” could be linked to SWEETNESS): what is important is that at any point in your research/discovery you can switch between the organization model to the user model.
Having popularity of product features is great. Having them distributed over a timeline and mixed with different demographic elements is even greater. However it is not enough. Popularity can be negative or positive so popularity by itself does not help to take decisions. VoU distinguishes positive from negative popularity which make the research of answer to product development questions much easier.
But that’s not all. There are some user attitudes that are crucial to understand market acceptance of your future product. These are very close to the classical “seven buyer motivation”, out of which we retained the following one: need/problem, greed, pleasure, intention to buy, perceived lack of features. VoU looks for linguistic cues characterizing these attitudes and relates them both to product type and relevant features.
Product types, product features, user attitudes, user demographics, timelines! At the end of the day you will end up with quite a bunch of information. Can you transform it into actionable knowledge (i.e. answers to questions posed during the whole product development cycle) ?
VoU optimize your experience by providing two kind of visualization:
just the graphics and figures you need to find, for instance:
which aspects of your new product deserve your special attention?
What is the “intention to buy” trend and by which features is it triggered?
How does your product relate to the age, sex, social activity of the user?
Which are the words to be used by the marketing to promote it?
An extremely flexible reporting platform (based on Kibana) that allows you to design flexible dashboards with multiple visualizations and tables. With few hours of training, this modality let the operator design powerful filters on data (views), custom queries, intersection, aggregation etc. In short DesignVoU delivers you the power of a Business Intelligence platform without the complexity which goes with it.
Sometimes product innovation is so disruptive that it is difficult to find product/services comparable with what your are going to offer. Sometimes the feature you are planning to propose was absent from any product comparable to yours. Sometimes you just need to go beyond features and attitudes detection to provide answers about your future product.
To handle these cases VoU offers a special functionality, DiscoveryVoU, by which you can track and generalize any kind of user expression.
Suppose you are questioning your team to know if it is worth investing on a revolutionary organic product that divides by two the water needs of home grown flowers. There is nothing like this on the market, how could you listen to the voice of the user? Thanks to DiscoveryVoU you can start with a small set of patterns identifying watering events as well as events where the user complains about wilting plants due to infrequent watering. Something as simple as “underwatered” “not watering enough” “dry ground” etc. The system will propose you ways of expanding these patterns (e.g. “poor irrigation”, “insufficient irrigation”) on the basis of your sources. After some iterations you will be able to provide a clear answer to your question: for instance by reporting that in 30% of cases of wilting plants the user speaks about watering.
The technical idea behind DiscoveryVoU is the one of democratizing a discipline which was long time developed in University laboratories, namely information extraction. Thanks to the integration of technologies from the Allen Institute for Artificial Intelligence with recent deep learning technologies, innoradiant is able to propose a solution by which any user is able to act as a text mining expert after only few hours of training.