She's right that ROI is not accurately or adequately measured for most SEO efforts, but her approach seems to be: well, if we could measure revenue from these other effects – branding, uplift in conversion rate – then all would be well. This seems a common attitude among marketers and I think it's wrong.

SEO is building an asset. SEO results in a website which in some cases can be worth a good deal of money. The value of the asset/website is not included in the ROI calculation, which only accounts for cash flows over a set period of time. The analogy is building a house for $100,000 which you can rent out for $10,000 per year. You don't look at these numbers and conclude your ROI is negative 90%, do you? No, because in addition to the yearly rent, you have the value of the house, which is set by the market and is probably a lot more than $10,000.

ROI is a better, if imperfect, tool for evaluating advertising campaigns, such as pay-per-click campaigns. For PPC, the residual value after you shut off the ads is usually small and can often be ignored. For SEO, even if you stop, the residual value can be large. Estimating the value of a website is a major challenge in itself, but that doesn't mean you should fall back on ROI just because you only have hard numbers for an ROI calculation.

ROI seems to be used only by marketers. You never see the rest of the business world using ROI. People in finance, accounting, manufacturing, real estate – they all use quantitative metrics, but they never use ROI. I suspect this is due to the fact that marketers have traditionally been people with poor quantitative skills and ROI is easy to figure. We never adopted the more sophisticated and accurate metrics needed to make good decisions. It's something the marketing profession needs to get over.