Articles by Nina Litovsky

Biases in marketing

Overestimating causality

Imagine that you are comparing your old data, and you see that in 2014 you averaged 10 new patients per month. In 2015, the only marketing adjustment you made was building a new website, and then you were averaging 15 new patients per month. You are tempted to assume that the gain of 5 NPs per month resulted from your renewed website, and you credit your web company for bringing in new patients.

However, correlation does not equal causation. It’s impossible to measure how much, in any, increase in ROI has been due to your new website. No one knows whether the increase of 5 NPs per month came from the new website or from other factors which may even be random and entirely unrelated to marketing. You may even see a decrease in NPs later, even though your new website will still be there.

It is useful to track plain statistical data such as the number of calls, visits, clicks, or form submissions. However, we should not twist the data and say that clicks and visits bring about conversions. We should not make a mistake of assuming that because we did X in marketing, it resulted in Y% of our ROI increase. Correlation between marketing and ROI is more causal and random than we would like to admit.

Survivorship bias

I see advertisements from marketing companies claiming that their methods are successful because some dentists followed their strategy and saw the increase of X NPs per month. These companies have case studies showing that their strategy worked for other dentists, which implies that if you hire them, they can likewise devise a strategy that will work for you.

However, a few case studies don’t prove anything. The results have to work for thousands of dentists to have any statistical significance. Hardly any marketing expert can claim that their strategy has worked for thousands of clients.

Besides, we don’t know how many of their (former) clients followed the prescribed strategy and failed. The few successful case studies that marketing companies feature in their portfolio is by no means false advertising – these success stories may have indeed happened – but it’s a clear example of survivorship bias.

What kind of marketing works for you?

There is no reliable, deterministic connection between a marketing strategy and revenue increases. Nobody (not even the marketing experts) can know for sure what works in marketing and what doesn’t. What works for one practice will not work for another. What has been working for you for many years may stop working at a certain point, and you will have to try something else. Also, what works for one segment of your prospective patients may not work for others. For example, some patients like an office with fancy decor and amenities, so you can advertise your fancy office to attract a certain segment of patients. However, you must be careful because other patients will be turned away because they may assume that you are too expensive. Some patients worry only about getting cheap deals; some others only care about the reviews and the professional qualities of the dentist. Even if your market is relatively homogeneous, individual people are looking for different things, so sticking to only one or two marketing messages is unwise.

You have to always check what works and what doesn’t, which means that you have to experiment and adjust your marketing strategies throughout the course of your business career. You also have to keep in mind that: a) some things, such as blogging and guest posting, are very effective and basically free, but it may take a long time before you see any results/traffic; b) on the other hand, some of your short-term gain that you attribute to marketing may result from other factors unrelated to marketing. For example, the number of NPs may not be caused by your marketing efforts, even though there seems to be correlation.

I don’t think there is a way for any marketing expert to guarantee that a certain strategy or a campaign will work for your specific practice, even if that expert has success stories in their portfolio and even if these methods work for other dentists. What will make your marketing strategy more robust is diversifying into many different directions to cover a variety of opportunities. Pick the marketing venues that are cost effective, that is, that can potentially yield large visibility and that, at the same time, come at a relatively low cost (in relation to your specific marketing budget). Experiment, and keep trying many different things all the time.


To illustrate this point, here’s an example. I build websites, but I wouldn’t say that my websites are necessarily “effective” or “converting” or “generating patient flow.” This is because correlation between a website and patient conversion is very causal and not deterministic, so I do not want to overestimate causality.

Now, it doesn’t mean that you don’t need a website. You need it because it’s a relatively inexpensive medium which can potentially yield high visibility and which will increase the chances of people finding you.

So I would not sell my websites as effective or converting, but I can build a website which is efficient (is nicely designed, readable, fast-loading, well-organized, has features that save people time and present information clearly and efficiently thus helping them to make a decision quicker, etc.) and which can make it easier for people to convert (such as by placing calls to action in appropriate places in relation to your marketable content, having related information where necessary, etc.). This is, by the way, why you need a custom website, and a custom website is not any more expensive than an expensive templated one.

So you need a website, but you need to have realistic expectations. Don’t expect a website to generate you patients, though a well-designed website can generally increase the chances of people finding you and converting.

Do not let anyone dazzle you by saying that the websites they build are effective at converting or statistically proven to convert. Tracking stats can show clicks and visits, but it’s impossible to reliably correlate stats with conversions. Also, even if their other websites showed good stats for other dentists (survivorship bias), it doesn’t mean their website will generate patients for you. And even if you get yourself a website that is advertised as “effective” and your stats show a bump in traffic, it doesn’t mean it will stay that way and not decrease over a couple of years. There are no statistically proven ‘secrets’ to make a website convert. Things have to be tested on many thousands websites in order to have any statistical significance. In website design, there are only a handful of common best practices that have been tested on millions of people and therefore widely accepted, and any good web designer knows these best practices. Other than that, I would be highly suspicious of anything that’s claimed to be statistically proven to work and convert patients.