How MakeTime changes the way you work and saves you time!

MakeTime^© was created with a vision of improving the way real people, in real life, access information.To improve on something we must first understand it. The way we search for information on our computers is MakeTime’s area of interest and expertise.

How Real People Search and The MakeTime^© vision

We believe that human search behavior can be broken down into a series of iterative predictions, evaluations and adaptations.

Let’s take a look at what we usually go though in a traditional search page:

The first step – prediction – is to enter, in a search box some text that we expect will eventually return the result we are looking for;

In a traditional search system we then click a “search” button and receive a page of results.

We then evaluate by visually scanning the list of results. In many cases we click individual results to further “drill-down” and examine the full page of information, before we can determine it in fact fits our needs. Often, several back-and-forth clicks are necessary to determine if we have what we want.

We then adapt our original search query (our original prediction) by refining and changing what we initially entered.

We then repeat – after clicking “search” we evaluate the new set of result, adapt again and evaluate again.

The process is repeated until we either find what we want or give up and possibly try again with a new query or go to a different search screen or system.

We consider abandonment a failure – in other words we fail when the search system was unable to lead us to the result we needed. This means that searching fails, when we are unable to express our understanding of what we are looking for in a way that allows the system to return the results that match our need.

Unfortunately, in most cases of failure, it is what we put in that is the root cause for not finding what we are after – this is usually because we rarely know exactly how the information is structured, organized and classified, what keywords or tags maybe assigned.

In other words, search is no different from any other computer or data processing system – we have the same old GIGO (Garbage-In = Garbage-Out) problem.

This is in fact a very serious problem, which is almost impossible to solve well with pure machine intelligence. This is one of those rare cases where the scientific method seems to fail to provide a framework for improvement. Our combined efforts as computer, linguistic, natural language processing, information retrieval and machine intelligence scientists and engineers are confounded by a chicken-and-egg problem – we cannot truly optimize a system for “the best” information retrieval, unless we have an unchanging set of inputs that we are going to design for.

Unfortunately, no two people are likely to express themselves in exactly the same manner on the same topic. Even the same person is likely to describe the same thing differently on two different occasions.

Still, over the last 30 years and especially since the internet changed our lives, an astounding amount of work has gone into improving search systems with the goal that they return the most relevant first page of results.

Stunningly, current research has shown that in many cases, incremental improvements of first page results relevance has no correlation to the ability of actual people to find information that they are looking for.

MakeTime approaches the problem differently – with a focus on the needs of the person, interacting with the search system, and not by simply pumping in more and more sophisticated algorithms. We aim to optimize the “predict->evaluate->optimize” cycle. Removing the search button and providing instant preview shortens the delay between prediction and evaluation. We dedicate as much screen real estate as we can to preview so that the evaluation can occur without back-and-forth clicking and with as much peripheral vision feedback as possible. Giving the user instant feedback with each keystroke and a constantly shrinking set of results provides the positive feedback that discourages abandonment.

MakeTime‘s contrarian approach to search is to discount the value of what many tout as must-have search algorithms and instead make sure that the user understands, at every key stroke, what the result is and, even more importantly, how they arrived at this result.

For example^, MakeTime is indeed capable of synonym expansion, pattern matching (single and double omission, substitution and inversion token errors), stop word filtering, statistical clustering, guided navigation, facet generation and a slew of other auto-magical methods of attempting to optimize the “Top-10” result list. We plan to offer them as standard features as soon as we invent a user interface that makes them self describing, self-explanatory and “natural” for the end user.

Today we would like to invite you to test drive MakeTime and see how we have merged the capabilities of full-text search with traditional parametric database sorting, range searching and individual field filtering. We aimed to eliminate the usability distinction between structured and unstructured information retrieval and tried to merge the utilitarian, “simple one-box search” with the often dreadfully inadequate, yet always lurking, “advanced search” form.