A common subject of my consulting is naming, and specifically naming the category of product or technology something goes in. Clients are well aware that no market categorization is ever precise. Still, words must be chosen, collateral must be prepared, and talks must be given to rapturous* audiences. Here are some of my go-to techniques.
1. My most precise tip starts from a classic naming dilemma:
- If we call it something entirely new, nobody will know what we’re talking about, or why they should care …
- … but if we say it fits in an old category, then how do we differentiate it?
Increasingly, my advice is to pick a name that’s “half new”, usually in the form of a two-word phrase that overlaps partially with the name of an old product category the new thing sort of resembles.
In some examples from my own work:
- ClearStory Data emphatically does not want its service called “business intelligence”, as doing so might denigrate the novelty of ClearStory’s innovations. A big part of how ClearStory beats traditional BI is in the intelligence of its data handling. Voila! With a nice bit of double-meaning, ClearStory’s secret sauce is now described as “data intelligence“. Edit (May 2014): “Data intelligence” has held up as ClearStory’s top message.
- Platfora’s latest release focused on data sets that — after Platfora assembles them for you — are sort of like time series but also somewhat like event streams. “Event series” was the winning name. Edit (May 2014): Platfora reports that that choice worked out well.
- Tokutek, whose main differentiation is performance, makes storage engines for MySQL and quasi-storage-engines for MongoDB. What should we call them? “Performance engines” fits the bill, at least for the “not exactly a storage engine” MongoDB case. Edit (May 2014): Tokutek hasn’t really stuck with that term.
2. A principle underlying that tip is that connotation is as important as denotation. The reactions that category names evoke can be as important as their literal meanings, especially since those literal meanings aren’t very precise anyway.
Returning to the examples above:
- I vetoed “event stream” for what we wound up calling “event series”, because “event stream” sounds like it has to do with super-low-latency CEP. It just does. That books have been written explaining otherwise is irrelevant; if you use that phrase, people will think you’re promising that kind of speed.
- When startups run away from the term “business intelligence”, they have a point. BI sounds like something one does against a relational data warehouse. Even so …
- … “intelligence” sounds eminently desirable. Everybody wants to be smart and to have useful information.
- Everybody wants “performance” too. As long as it doesn’t convey “expensive” or “troublesome to use”, it’s a great term.
Similarly, I once renamed graph-based “relational” analytics to “relationship”. Calling something “relational” is counterproductive when the main point is that it doesn’t involve a relational DBMS.
3. The naming task isn’t quite as hopeless as it might seem, because names don’t stand alone. You always put at least a couple sentences around a name, and that helps a lot with offsetting certain kinds of drawback. In particular:
- While a two-word phrase is likely to be vague, two sentences can make it more precise.
- A short name will rarely convey any vertical focus, but an additional clause or two can usually close the gap.
4. Even so, take great care that you don’t convey a clearly wrong impression. This is of course a general rule of marcom — except for a certain amount of generally-acceptable puffery, false communication is always bad. But the rule is particularly applicable to naming. If you place yourself in the wrong category, you’re unlikely to wind up in the right deals.
5. Combining the points above, too vague is safer than too precise. E.g., part of the reason the new term is “entity-centric” event series analytics rather than “user-centric” is that, while most of the entities in question — whatever an “entity” is — are users, not all are. The other part of the reason is that “user” just sounds like it has something to do with UI, or with the users of the vendors’ technology; hence it doesn’t immediately sound like it’s about analyzing the vendors’ customers’ users (or user-like entitities).
6. At the risk of redundancy, I’ll close with a Hippocratic precept: Above all, do no harm. The best name is unlikely to do you all that much good. You have to explain it, and support the messages it conveys. But a bad name can cause a great deal of trouble, getting you into deals you will lose and keeping you out of ones you could win.
- I covered positioning more generally last April.
- The entity-centric terminology (vs. user-centric) actually started with my clients at WibiData.
- Notwithstanding their own preferences, ClearStory and Platfora were both featured in my recent post on specialty business intelligence.
- And finally, Ursula Le Guin’s story The Rule of Names has no legitimate connection to the subject of this post. But it’s such a classic that I’ll mention it even so — a short, evocative and even funny piece of fantasy that foreshadowed the magic system of her Earthsea Trilogy.