Guest Post: There’s no such thing as MEDIUM

Add to FacebookAdd to DiggAdd to Del.icio.usAdd to StumbleuponAdd to RedditAdd to BlinklistAdd to TwitterAdd to TechnoratiAdd to FurlAdd to Newsvine

NOTE: The following is a guest post by Tzvika Barenholz, a Product Manager living and working in Israel. If you feel inspired to write a guest post of your own, click here to find out how to submit it to us.

We all know the situation: you’ve collected a bunch of product requirements, put them into a nice excel sheet or clever table-like GUI front-end; you’re all psyched about making the big decisions, prioritizing and leading the product to the proverbial “next level”. You sit in front of the screen, crack your fingers, stretch a bit and place your hands in the typing position.

“First item, feature xyz” – yes, you say it out loud, you’re psyched remember? – “First column, Importance” (or priority, or appeal, or whatever is your favorite alias for customer value)

OK – now what *was* the priority of feature xyz?

Let’s see. A couple of big customers have asked about it in a recent expo, so probably high, right? Then again, it’s one of those things that make the product more complicated, so call it low. Having said that, it would really improve average sale price, which is aligned with the five year plan, so high it must be. But what about breadth? It really would only appeal to a fraction of our customers. Back to low again.

And on, and on.

Exasperated, you note that after 20 minutes you’re still on item #1 out of 37. It’s looking more and more like a long night of pizza and coke with the developers. The next thing you know, there’s a little voice inside your head that whispers: make it a medium!

Of course! How foolish you’ve been to overlook this possibility before. A medium is right in the middle between high and low. It’s the perfect answer for what seems to be neither here nor there, or both here and there. Your right index finger then makes it merry way to the m key.

But STOP – you’re just a moment away from falling into a trap as old as the hills. Why? Because there is absolutely positively no such thing as a medium, except maybe as a shirt size.

“Medium” is a shirt size, not a useful priority rating

Prioritizing something as medium is just a way chickening out of making a decision. If you don’t have enough information to make the call, go get the information, then make the call. If you’re piling up different qualities like “broad appeal”, “helps sales”, “improves margins” and “technologically strategic”, then by all means, add 4 columns, fill them out with highs and lows (or yes or no, for the more advanced practitioners who know how to make an even cleaner cut), and then rebuild the priority column as an index composed of them all.

But whatever you do, don’t just cop out. Don’t just type medium and move on to the next item, or you will end up building the wrong things for the wrong reasons.


11 responses to “Guest Post: There’s no such thing as MEDIUM

  1. Medium is when you move from zero to one. It’s that period of indeterminacy between doing nothing and doing it all.

    How many times have we had to ship before we did it all? We ended up shipping what we had. Sure it was only some of the all we wanted, but it was something. Heck, it could have been nothing.

    Medium is the story we’d rather not tell about ourselves.

  2. Your 4-column method is the beginning of a simple decision matrix. This can be as simple as you describe or become as complicated as a QFD/House of Quality chart. One important factor that also needs to be included in any prioritization exercise is difficulty. For an illustration of why, see this short YouTube video – the Product Management Trap.

    Other references:

  3. ‘Medium’ is the ‘don’t know’ that you revisit on a second pass once you’ve figured out the bony, essential ‘must-haves’ and need to flesh out the release plan so there’s a consistent theme.

    Medium selections get driven by your organization’s strategic product aims, whether they are technical, market-focused or commercial. This is where you as a product manager get the most leeway and exercise your professional judgement to implement the direction your product will take.

  4. I completely agree. When you look at priorities from a release perspective, High gets done, nothing else does. Medium has no bearing. We’ve been trying to create a composite priority system as described in the post by rating features against several corporate priorities, weighting them, then combining them into a single score. after several iterations, I’m still not completely satisfied with the results. For those out there that use scoring systems to prioritize features, what different factors (in addition to the ones in the article) do you score?

    • John,

      I have a little saying that I often use. “We’re here to make decisions, not calculations.” 🙂

      Decisions are always made with incomplete data and thus have inherent uncertainty. There’s a place for good research data, but it is in support of the decision making process. Unfortunately, many companies generate artificial data by creating ratings systems and use those metrics as justifications for decisions.

      I’m not against rating systems, as they do help put focus on issues, but I view them as a coarse grain tool to *help* cluster issues or topics that deserve more focus or to eliminate topics that are a poor fit.

      My advice on your rating system is find the 3-5(max) categories such as strategic fit, competitive impact, customer/partner pain, etc. that make sense in your context and use those as guiding factors.

  5. Tzvika Barenholz


    Great video and links! Personally, I like to “Keep It Short & Simple”, so I usually opt for the simple matrix, and stay away from the full-blown QFD. The point of my post, translated into the 4 quadrant system, is: “don’t put anything on the lines between the quadrants”.

    Owen & John:

    OK, fair enough, you don’t know in the first pass? leave it empty, come back to it later. But in my experience sometimes what John said ends up happening, and only the first round of Highs gets scoped. So it amounts to the same thing, as calling it low to begin with, or fighting for it as High. Either way, the earlier you make the tough call, the likelier you are to get it done.

    Thanks for the discussion, all!

  6. I tend to go for stack ranking myself. If I need a higher level prioritization, I tend to create revenue/cost buckets (e.g. Priority 1 = committed revenue, Priority 2 = likely revenue, Priority 3 = no revenue impact but likely cost reduction).

    • Tzvika Barenholz

      Tim ,

      I like the idea of stack ranking, and I’ve seen it work on occasion. However, in practice sometimes I’ve ran into a brick wall using it for two reasons: One, when you have 100 items, entering a new one becomes a rather messy process of comparing it to many items before you can figure where it fits. Two, when you have different dev resources working on the same stuff, each item requires attention from a subset, and it then becomes quite possible for things lower down the stack to happen before more important stuff, simply because they’re easier or do not require some bottleneck resource to happen. It can somtimes turn into a “path of least resistance” system.

      But I guess that’s true of any method.

  7. Saeed,

    I agree with your comments back to John Westerveld, I’ve implemented a feature / requirements rating system for a number of clients and the biggest benefit they all said was that it gave them focus. The advice I give every client is to take a step back and review the results of your rating exercise and ask if it makes sense based on the information at hand. The key in any rating system is the objectives against which you do the rating … they need to be aligned with the corporate objectives and be at an operational level.

  8. Pingback: Twitter Trackbacks for Guest Post: There’s no such thing as MEDIUM « On Product Management [] on

  9. There’s another aspect of this discussion, which is the weighting you give the different columns. Well, there’s no one answer to this. For one quarter’s release you might be focused on satisfying a particular customer. For another quarter, you might focus on opening a new market. Those strategies will result in quite different requirements being prioritized up. And it’s very hard to do that either with a spreadsheet or in your heard (as in stack ranking).

    I happen to have a tool that does this for me – and it *does* work well. (Unfortunately, it still hasn’t solved the physics problem of delivering all those High priority features with limited resources!)