This is exactly what I'm aiming for in developing Instahero (http://www.instahero.com). Your traditional analytics software has all this data, why can't you get a metric you need right away?
Instahero lets you just write one or two lines of Lua and visualize/report on any metric you want.
Sorry about the plug, this tool is just similar to what I'm working on to solve a more general problem.
Instahero looks cool. I was using geckoboard & I though pg's suggestion of focussing on a single metric made a lot of sense, given the amount of time I see startups spend on custom metrics of all colors & stripes.
There's definitely a tendency for people to focus on vanity metrics, so pg's way makes a lot of sense. My post here is a bit orthogonal, I guess, since pg's essay and your tool are about what to measure, whereas my post is about how.
That said, I do like the tool, it's very easy and straightforward, good job!
Maybe you can take the idea & turn it into something for instahero? Think "landing pages" for certain types of metrics pre-configured. instahero/weekly-visits | instahero/weekly-subscribers | instahero/weekly-visits | instahero/weekly-revenue etc ...
I like the idea of having a single page per metric
This is very easy to create using whichever metric you like (I will probably include it as one of the pre-made report templates, too), and you can extend it to report on whatever you like.
Should the JSON returned be the set of weekly numbers or daily numbers? It doesn't say and since you mention that it's going to be crawled daily it seems like it would be daily numbers...
"Are you measuring your weekly growth?" No, and for very good reason. For pete's sake, you should not be measuring your growth numerically and targeting it weekly. This is not what Paul meant, and I don't think it's what he would recommend.
Growth happens on a larger and more complex scale than a weekly sum. Furthermore, if you're expecting your actions for any given week to correlate directly with growth that happens that week, you're doing it wrong.
Do you need to work hard? Of course. Should you procrastinate? Of course not. Is the key to this tracking your weekly growth. No! You risk misdirecting your efforts, becoming far too short-sighted, and falling into the dangerous trap of running your company only on those quantifiable and measurable numbers. You risk missing the complexity which results in growth, and instead narrowing your focus to the point where it actually stagnates growth or produces the wrong kind of growth.
See W. Edwards Deming's management treatise that helped bring Japan out of a recession—http://en.wikipedia.org/wiki/W._Edwards_Deming#Key_principle... - one of his recommendations was to "Eliminate management by objective. Eliminate management by numbers and numerical goals. Instead substitute with leadership." And a pitfall he pointed out—"Running a company on visible figures alone." More can be found in his books, which mostly are targeted toward manufacturing but I find apply incredibly well to software business.
Please, don't limit yourself to this sort of narrow-minded analysis of your business, or at least be very aware of the consequences and consider it only as an interesting result, never use it as a target or benchmark, and especially don't use it to try to motivate or pressure your employees. Your results will slowly shift toward optimizing your measured goal alone, and your company will become as narrow and pointless as the number.
Can this metric be used? Of course. You can look back on your efforts over the long (or short) term, and if you see a decent 6% growth rate, great. If you don't, then you may need to change something. But you have to understand it intimately; that it's not a goal, not a target, but a simple metric to be used post-facto to make useful and realistic changes to your strategy. I'm not warning you against using all numeric analysis, but I am warning against using it as a true goal or target. Just keep it in perspective, and I fear a web site that makes the metric as clear and visible as this site does brings it far too close to the front of your daily attention than it should be.
Growth comes from many complex sources, toward many time frames and overarching goals; and the target is not a weekly number, but instead, a growing and sustainable company. Work toward that.
PG: We usually advise startups to pick a growth rate they think they can hit, and then just try to hit it every week. The key word here is "just." If they decide to grow at 7% a week and they hit that number, they're successful for that week. There's nothing more they need to do. But if they don't hit it, they've failed in the only thing that mattered, and should be correspondingly alarmed. </unquote>
Maybe the caveat should be: "If you want to play the high-growth startup game". Measuring weekly growth apparently is a tool that works in that context.
Perhaps I disagree with PG. I stand by my post. However you're correct, in that context it could be a useful metric when the goals are very focused, the targets are well defined, and the growth is directly correlated. He could be entirely right in this context, you're right.
It is my opinion that this sort of thinking builds unsustainable companies. There either needs to be an incredible shift in the way growth is looked at and achieved at some point in the company's life (which to me, seems unlikely or at least a difficult change of course)—or the company can start out with sustainable growth in mind. There is no reason that a company designed for overall quality rather than speed cannot also be a fast growing company; but I think they'll look at their strategy much differently. And that way of thinking will last them through hiring, market expansion, and beyond.
Maybe they are compatible strategies and I'm just over-thinking it. But I can't help but see the recent articles about how few long-term successful companies arise from tech accelerators and think how strikingly relevant they are.
I tried to keep the6pct sparse on purpose. I think it's the focus that helps. But you're right, this should come with a pharma-sized disclaimer & I'm afraid I didn't do a great job at that, not even in the accompanying blog post:
Ah, context: "you should not be measuring your growth numerically and targeting it weekly." My point is not that the measurement is numeric (of course whether it's numeric or not is irrelevant); it is that using a sum number on a weekly scale can limit your view, narrow your goals, and have the opposite of the intended effect.
> This is not what Paul meant, and I don't think it's what he would recommend.
"During Y Combinator we measure growth rate per week, partly because there is so little time before Demo Day, and partly because startups early on need frequent feedback from their users to tweak what they're doing. [..] This is, obviously, only for startups that have already launched or can launch during YC. A startup building a new database will probably not do that. On the other hand, launching something small and then using growth rate as evolutionary pressure is such a valuable technique that any company that could start this way probably should.
"
"using growth rate as evolutionary pressure is such a valuable technique that any company that could start this way probably should." -> that looks like exactly what the 6% check is for - a simple tool to keep you motivated, not to rely your whole strategy upon.
Instahero lets you just write one or two lines of Lua and visualize/report on any metric you want.
Sorry about the plug, this tool is just similar to what I'm working on to solve a more general problem.