Unstructured data is notoriously difficult to structure and query, but Impira Query Language (IQL) enables users to get “What” they need from information systems without worrying about “How” to search for it.
When we want to do something we haven’t done before, the “How” is usually what holds us back. How do I measure my business’s success? How do I plan for retirement? How do I make those Ram-Don noodles from Parasite? With all these questions, you know “What” you want, but not “How” to get there.
So what do you do when faced with a question like this? You probably just type the “What” into your favorite search engine and get an answer around the “How.” More and more services these days are following this “What, Not How” philosophy. You as the user just describe the “What” and leave it to your chosen service to figure out the “How.” But “How” is all this possible, technologically?
Solving the “How” comes down to letting users access data in efficient ways using the “What.” Historically, relational databases were among the first solutions to this problem. The arrival of SQL allowed users to express “What” they wanted declaratively, but abstracted away the imperative “How” (i.e. the technical procedures involved in accessing and transforming the data). For instance, consider the SQL query ‘SELECT id FROM users WHERE age > 25.’ Even if you didn’t know SQL, you could figure out that this query is asking for information about users older than 25. It’s almost English!
When using SQL, you don’t have to worry about which algorithm is used, how data is retrieved, where it’s stored, how many computers it’s stored on, etc. It just works.
Fast forward to today, and despite continuous advances in database technology, companies continue to struggle with the “How” of accessing their data, especially unstructured data scattered across cloud drives and databases. They know the “What”, but don’t know “How” to query across unstructured data—or even “Where” to query for it in the first place.
Data comes in all shapes and sizes and lives in different locations. For instance, companies may have tabular data in SQL databases, but also image data in Dropbox. Data in one location may point to a different form of data in another location (e.g. URLs in a spreadsheet pointing to images). Further complexity arises when searching over this data, as you may need to refer to a field in one location and then follow it to another.
Today, the average medium-to-large enterprise employs a dizzying array of homemade and third-party systems to manage its data, often with considerable overlap between these systems. At larger companies especially, different lines of business (LOBs) have the autonomy to select their own systems, which results in “application sprawl.” Salespeople might house visuals in a PIM (Product Information Management) while marketing people house them in a DAM (Digital Asset Management), with little integration between these systems to ensure brand consistency, updated versioning, and easy exchange.
The “Where” problem tends to manifest even at smaller companies, where application sprawl is less of a problem. We hear again and again from existing and prospective customers about the challenges that accompany disorderly cloud drives—lost assets, inconsistent keywords, and folders nesting to infinity. Given the superabundance of data piling up in every organization, even users who know exactly what data they need often have a hard time finding it.
Clearly, we need a way to bypass the “Where” and the “How” in order to arrive directly at the “What.”
In an ideal world, users should be able to find “What” they’re looking for whether or not they have all the “How” information at hand. Not knowing the storage location or exact naming convention of a given image shouldn’t prevent you from finding it. Likewise, users with only an inkling of what they’re looking for should be empowered to go on a journey of discovery.
The searchability issues plaguing organizations today are in large part what inspired us to build the next generation of “What, Not How” querying: Impira Query Language (IQL). IQL empowers users to search across silos and data types using dynamic filters and natural language (not unlike what you might use in a search engine). For example, you could search for a product photo “sku__123__henley__male.png” with a query like “sku: 123 henley” or "henley male" or even just “henley.” Here, we combine free text (“henley,” "male") and dynamic filters (“sku: 123”) to bring users a search language customized to their vocabulary.
This language is compatible across different data sources and backends, but all this is abstracted away from the user. IQL is infinitely flexible in allowing users to search by the parameters most relevant to their day-to-day needs and engage in an exploratory journey with their data.
For our users, IQL means never worrying about the “How.” Impira will find your data for you, so you don’t need to worry about where it is or what form it is in. Once connected to your data sources, we provide structure to hard-to-parse data and index it automatically—granting you peace of mind whether or not you know exactly “What” you’re looking for.