This is one of the things I love about working on SportHold as a startup.
If I were in a corporate role and wanted to use my digital marketing skills to boost traffic I would have had to ask for approval, create a proposal, discuss it with managers, etc...
I'd then have to broaden it to cater to the whims of others involved and the diluted version would suck. The experiment would fail and I'd be held responsible. The lesson would be that, categorically, we've tried X and it never works. Proposing X again in the future would be a lost cause.
But in startups I'm free to do what I know and have experience in. I'm able to quickly execute experiments, learn and evolve. The big bump is the experiment phase where several things worked, some didn't and some worked but weren't cost effective.
The decline after the bump is where we dropped the ones not working and most of the ones not cost effective. We then ramped up the cost effective ones (cost is not necessarily money but also time and energy) and looked at extending the experiment for the ones that weren't cost effective but we felt could be.
We continue to test and experiment. Since the decline after the bump we've been able to double our daily traffic & user numbers with marginal increase in cost.
But it's not just vanity metrics. Three weeks ago we had 1,700 users. Today we have 3,874.
For the things that didn't work (this time) they're strategies we'll be open to revisiting later on when the context has changed. A failed experiment doesn't have to have far-reaching implications about overall digital strategy because startups are fine living with uncertainty and unknowns.
We don't have to force conclusions that aren't there because startups don't have a CYA culture that demands black and white judgments.
The best part is that we've just barely scratched the surface of what we can do in terms of digital marketing.
UPDATE 13th of May, 2014: Here's what the chart looks like now: