Tom Fishburne nails it again:
As he says in the accompanying post, the “waterfall” approach is a terrible fit for web project management:
We couldn’t predict all of the potential issues when we first wrote a brief. So, requirements would inevitably change and we’d uncover issues too late to do anything about them. We would then either stubbornly hold to our original plan or scrap a lot of hard work to start parts of the process over.
Waterfall development also stifled great ideas that came mid-stream. It clamped down on “feature creep” to the expense of creating more remarkable innovation.
That’s not to say that no effort should be made to capture requirements up-front. Certainly there are guiding principles that are understood about many web projects from the beginning, but those are often business requirements and/or user requirements, not functional requirements. Trying to dictate functional requirements too early in the process often leads to a product that misses the mark entirely with features no one uses and a gap where the killer features ought to be.
The relevance this has to marketing optimization is that this is the scenario we frequently encounter when we meet with new clients and prospects: their web site and all of its associated marketing programs have been created under a waterfall-like approach that has its basis in an outdated understanding of requirements. Most aren’t testing, and those that are aren’t often processing the findings from their tests at a strategic business level.
And believe me, I say this not to dog our clients. I have the utmost respect for our clients: they’ve recognized the need to do something more progressive than redesign and rebuild, and thus start the whole vicious cycle over again.
It’s the businesses who don’t recognize the need for change that most need it. Take that waterfall diagram, and think about your last few web projects. What if you could correct your course along the way? What if you could find the parts of your web presence and your marketing strategy that aren’t working as they should, WITHOUT having to redo everything? What if, when you do decide to redo everything, you could base your decisions on results from what you’d already put to the test?
This is the reality of optimization, and it just makes sense. Any change you make today is a change you could be testing, and by testing, you can know the good you’ve done in objective terms, or, if the change proves ineffective, know in an equally quantifiable way the harm you avoided by not making a sweeping change and wondering why you missed your targets for the year.
Not to say that there are no risks in testing, and no trade-offs: certainly the process can feel unnecessarily cumbersome and slow if you’re used to having carte blanche freedom to make changes to your marketing and your web site. This can be especially frustrating if your title has a “C” and an “O” in it, and/or you’re used to being the de facto creative director on all campaigns. Data can feel a little intrusive in those cases. But the dark side of making decisions based on HPPO (highest paid person’s opinion) is that you get held accountable (whether by your CEO, your board, your shareholders, or your customers) for data-driven results anyway, so why wouldn’t you use data in your decision-making?
Happy Monday. May your data be accessible and your findings be insightful.