In a moment like this, you have to *very* careful and stick to your integrity... because to be honest, there are a *wealth* of black and gray hat tactics that will "solve" the problem for any single metric. And if your client keeps hammering away on a fix, you might even be tempted to use them. But they only create, well, much bigger problems.
Open rates in email an issue? Change your sender name to something salacious, or tweak the subject line to oversell a benefit or personalization. Sure, you may lose in-boxing, get booted by your provider, and unsubscribe and bounce rates will spike, which are all way much worse problem to have, but hey... you only wanted to fix the open rate, right?
Let's go further down the funnel, then -- clicks are tried, true, and never high enough to make everyone completely thrilled. OK, dumb everything down to a single entry point (for old-time fans of the blog, the Jolly Candy-Like Button), make the call to action pop to the point of obnoxiousness, and prevent anyone involved from analyzing site behavior or caring about click quality. Bonus: show the ad in a new geographic area for the consumer category, even if it's impossible for them to make a purchase. Voila, CTR is spiking! Not conversions, though. Definitely not conversions.
Well, fine, Mr. Black Hat Consultant... let's only measure for conversions, then. That'll fraud-proof it! Except that in making this move, you've likely made testing impossible due to the rare event issue, made everything dependent on a conversion funnel that is likely independent from lead generation, and given everyone involved a massive incentive to sell on an irresponsible price point and forget about return on investment or lifetime consumer value. If you are only measuring on conversion, well, converting on a loss leader isn't really that much of a trick. (I'd get into SEO here, but the black hat work there is so prevalent, it's honestly hard to sell honest services.)
The point is this: life (and marketing) is often *complicated* and a complex problem to solve. Trade offs are inevitable. There are very few things where you only want to know one metric, and are ready to toss all others out the window. If the only thing you care about is how fast you are going when you drive, you're going to run out of gas. Or fail to heed the check engine light. Or drive through a red light while you stare at that speedometer.
Marketing and advertising is the same way. A long-term marketing professional wants to see as much actionable data as possible -- because they give us clues for how to make things better. Find out what site behavior says about click quality. Work out what platform someone is viewing your offer on, and how they index demographically. Grind away on small but free levers like dayparting, frequency, segmentation and so on, and so on. And when you measure for impact, look at more than one point in the funnel, if only to check that your gain in one place isn't being wiped out by a loss somewhere else.
Any marketing and advertising consulting agency should be able to help you solve your most pressing problem. But if they do so without making sure they aren't creating others, that's not good or sustainable service.
