3 ms·
My very rough and sloppy analysis gives you a credible interval of a conversion rate between 0.28% and 5.61%. At the upper bound of that conversion rate (5.5%),
by gburt 13y ago
My very rough and sloppy analysis gives you a credible interval of a conversion rate between 0.28% and 5.61%. At the upper bound of that conversion rate (5.5%), I would handwave and suggest this is comparable to your Google CPC results (depending on the actual cost per click). If I were you, I'd collect more data.
And try different ad text. Acknowledge that this is a different platform than search and you need to advertise differently. Don't be so quick to dismiss it.
Edit: and I was comparing apples to oranges anyway. If I use your Google Analytics data for both measures, we get a range of 0.39%-7.7%. This upper bound actually exceeds your Google CPC result. You don't have enough data.
- tinco 13y agoHow confident?
- gburt 13y agosig=0.05, but please don't trust that for anything meaningful. This is sloppy 8am blog post response statistics. My personal confidence is basically 0. I just wanted to communicate the idea that 60 impressions was insufficient to make this judgement.
- tinco 13y agoIf I'd do such a study, I'd pick a lower confidence interval, like 0.2. This is not a big scientific break through, or a business critical decision. This is a theorem that we're evaluating to be worthy of deeper investigation. If there's an 80% (also known as 'a significant') chance that the conversion rate is significantly lower than that of Google hits then I'd say that warrants deeper investigation. That's why it's important to always begin research with a hypothesis, not just randomly throw a confidence interval of a something in there :)