Not long ago, I was invited to participate at a customer’s annual conference. It was an amazing experience. I’ve been to conferences of all sorts, but I confess I’ve never attended an event quite like this one. Let’s just say that I’m used to.. well… less energetic IT conferences. This particular company is *extremely* good at marketing and really understands the power of hype. The combination of pounding dance music, an elaborate stage set up, spectacular lighting, and, most importantly, well crafted and super hyped product announcements had the 20,000+ attendees in a frenzy.
Now, before you start thinking that I just insulted this customer’s business by using the word “hype” twice in describing their event, understand that I mean it as a sincere complement. Hype, short for “hyperbole”, means “deliberate or extravagant exaggeration” and is a well established and ancient promotional technique. Let’s face it, marketing hype has become so fundamental to our attention economy that, with a nod to Joel Gray and Liza Minnelli, we might say that hype, not money, makes the world go ‘round. Successful businesses, like this customer of ours, know how to walk the fine line between powerful marketing messages that attract customers and ridiculous exaggerations that turn them off.
After I got back from that conference, I started thinking about hype and its use in enterprise search marketing. I looked at what my own company has produced in the way of marketing material and then took a look at some competitor’s sites. I found nothing particularly outrageous. Some of us are more prone to hyperbole in our marketing than others, but relative to other industries, we are pretty tame and rather typical for the IT industry I think.
Even so, I thought it would be interesting to look at the more common examples of the hyped up statements I found and to try to offer my own translations. An example of a translation looking something like this:
“He’s as big as a house!” (hyperbole)
“He’s a large man.” (translation)
So, here goes. The top 5 hyperbole statements in enterprise search marketing and my translations are below. I’ve paraphrased them, so don’t bother trying to do a Web search to find the sources. You may still find something if you do, but it’ll be a pure coincidence.
1. “Access Any Content Source”
We provide an application programming interface (API) that you can use to write content feeding applications for our platform. You can use this to submit documents, db records, or any other types of information objects in order to make them searchable, as long as the objects you send in comply with the APIs protocol and are in a format that our platform can recognize and translate.
We may also provide, directly or through partners, a set of content source “connectors” or “adaptors” to standards-based (e.g. Web, file system, database) and proprietary information (e.g. Lotus Notes, MSFT SharePoint) sources. Some of these connectors may work in a way that simplifies keeping the search engine in sync with the source. That is, they may be configured to periodically look for new, changed, and deleted records or documents in the source. Fewer of these connectors will work with the native update mechanism of the source system so that the search engine can be notified of additions, changes, and deletions the moment they happen. Lastly, still fewer of these connectors will respect and pass through to the search system the native access control information of the source so that searchers don’t see things they’re not supposed to see.
2. “Infinite Scalability”
Our enterprise search platform has been designed to scale, theoretically, to an infinite amount of content. This is because of an architecture that can be distributed across multiple machines and, therefore, can grow to include more capacity by adding hardware. You should not conclude, however, that it can scale up cheaply or easily. To get to very large volumes of content and queries, you may have to buy a lot of hardware and data center space. If the platform supports any features beyond very simple search, you may also find that turning on these different features impacts how effectively it scales – that is, what you can get out of your hardware investment. And it may or may not include consideration for scaling on both the content side and the query side, so if you grow the amount of content your system is serving, it’s possible your query capacity (the number of queries your system can handle within a unit of time) may actually decrease. (See previous post on search system scaling.)
3. “Access to Any File Type”
We have file format converters for a great many different file types, but then most vendors can legitimately claim support for more than >300 formats including all the major office suite document formats and associated versions. In practice, most enterprises with intranet search application requirements will care about only a dozen or so of these – Web or HTML/XML documents, Microsoft Office formats, Adobe Acrobat (PDF), and a few more.
Image file formats can be searched if they have associated meta data, or if you incorporate object character recognition (OCR) capabilities for scanned document images. The OCR feature may be provided directly or through a partner. Similarly, audio and video content may be searchable through meta data, or, if the ability to search full-text transcripts is desired but transcripts aren’t available, through speech-to-text conversion technology.
4. “Support for over XX languages”
We are very confident that our platform can handle the XX languages that use the standard character encoding sets we can handle, even if we maybe haven’t tested every one of those languages. However, more involved linguistic processing for things like synonyms, spell checking, entity extraction, and other advanced features may only be available for a small subset of languages and we may or may not provide 3rd party options or the tools to build these capabilities yourself for languages that we don’t cover.
5. “Language Independent”
(A variant on the last one, but worth its own translation, I think.)
The core of our platform is based on statistical model for calculating relevancy of items in a search result. Because it’s completely statistical, it is theoretically language independent. However, the quality of the search results is improved by, and some features depend on, knowing the language of the content and the query. For example, one of the most basic things a search engine might do is break the text in a document down to individual words and/or phrases in order to build an efficient retrieval model (index). Since the rules for how to do this “tokenization” vary by language (e.g. in Japanese, you can’t always rely on whitespace as a separator for words), it is helpful to have language specific extensions. So, while you may be able to use the system with any language, it may not work very well and your users who depend on that language may complain – bitterly.
Bonus – “100% Precision (or Accuracy)”
(Ok, I admit that I’ve only heard of a vendor making this particular claim once or twice and I have never seen it in writing, but it’s my personal favorite. I don’t know quite how to translate it, but maybe it’s like this… )
Our search software has God-like powers. It can not only read your mind and understand your information need beyond what you can comprehend yourself, but it has access to all known and unknown sources of information in the Universe.
There you have a few examples - ones you may see, or hear, in one form or another in enterprise search marketing (except for that last one, hopefully). If you know of other examples, please share them. My hope in providing these translations is that maybe they will give potential buyers something to think about when sifting through marketing literature.