Recommending a Guest

Since inception, Data Skeptic has doubled in listenership every few months. We are overwhelmed with gratitude for the response and the community that has formed around the show, particularly in our Slack channel.

Unfortunately, the growth of the show has also meant that we can no longer respond to all correspondence. We do our best, but when responding to all emails started taking more than an hour per day, we have to put some limits in place. We are always grateful to get emails from people saying how much they like the show, even if we don't reply. We also appreciate critiques, corrections, and feedback of any kind. Emails that have short, to the point, and ask a clear question are most likely to get a reply.

Lately, we've been somewhat overwhelmed with emails recommending guests for the show, and I wanted to specifically respond to these inquiries so that I could share this link in appropriate situations.

The suggestions we get are a mixed. From time to time, we get a passionate email from a genuine fan making a great suggestion of a colleague as a guest. Such personalized emails are often great ideas, and even if we can't accommodate, we very much appreciate people making these suggestions.

However, for every good suggestion, there are 19 others which are template emails from PR agencies recommending people that are rarely even close to a good match for the show. To help stem this tide, I wanted provide some feedback on our guest selection process.

First and foremost, we're moving towards having themes on the show. Listeners will notice this becoming more formal starting in October and for the foreseeable future. While we might include some one-off topics, we're going to start focusing all guests on a specific area for 2-3 month runs. Thus, regardless of how interesting a guest may be, we'll only be interested in people that fit the theme. Additionally, our production schedule now demands that we often plan far ahead, so many themes may have all guests booked before we star airing.

All of this is to say, we definitely expect that the demands of the show will mean we sometimes turn down suggestions of guests that are a great fit because they aren't congruent with our current themes and topics.

For PR agencies taking the time to read this post, below are a list of specific points about guests we are not at all interested in. Sending suggestions that violate several of these bullets will likely mean we don't even reply. We are not interested in guest suggestions that match these criteria:

• The solicit focuses on how much money a company has raised. We couldn't care less about funding on Data Skeptic and are unlikely to care about a guest whose primary talking point relates to funding over content.

• The suggestion says little or nothing about the specific topic the guest might speak to and only talks broadly about the categories they are allegedly qualified to talk about. Make sure the guest has a specific project, paper, or topic of some kind to focus the interview on.

• The main talking points in your email stress what VC firm someone is working with, where they worked previously, or how great of an entrepreneur they are. The word "visionary" is almost always a bad sign. We want to interview people who can speak about novel and interesting work they are currently engaged in or have recently finished.

• We are not interested in guests actively on a media tour, appearing on many media outlets. This makes their appearance on Data Skeptic redundant.

• Generally, we don't want to speak with executives. Executives are rarely able to speak about anything more than surface level details. Data Skeptic brings technical knowledge to our listeners. Before recommending an executive, consider instead suggesting someone in their organization who is doing the actual work. This extends to people with titles like "evangelist", "media coordinator", and "community developer". Data Skeptic is not a platform to advertise products and services. We want to talk to technical people who are doing the actual innovation.

• When a guest "can't talk about" certain topics, they're probably not a guest we're interested in. If the interesting details worth talking about are considered trade secrets or proprietary information, then they're not a good fit for Data Skeptic.

• If a guest is part of an organization that wants to have a review of the final edit, then this is a 100% show stopper. We've entertained this twice in the past and both times, it was a mistake. We have years of episodes that demonstrate the style and content of our interviews. You are free to review them to understand more about the types of questions we ask. If you feel our interview style is too invasive or might reveal details an organization would rather keep private, than we will always respectfully decline the opportunity to interview someone from that org.

• When the guest suggestion mis-represents the true expertise of the potential guest, not only won't we be interested in the guest, but we'll likely blacklist the agency making the suggestion. It's surprisingly common for PR companies to send emails that attempt to portray people with expertise in fields other than data science (for example cyber-security) as being "expert in machine learning". When a PR firm behaves deceptively in even one case, we assume they'll consistently be unethical, and thus, we don't want to work with them.

These points are shared to give guidance, but not to discourage. If you think you've got a great suggestion for us, we're eager to hear it. Our process is to place all suggestions in a queue which are reviewed on a bi-monthly basis by our editorial team. We'll definitely review all suggestions, even if we aren't able to reply.

However, one tactic that will get a suggestion removed from our queue is to make needless follow ups. Replies asking if we got an email are not appreciated. Worse still (this is a real case) a reply saying "just bumping this to the top of your inbox" are the quickest way to get your address blocked by us. We prioritize the best we can, and these tactics are disruptive to the point that we don't want to work with anyone that uses them.

So what do we look for in our guests? Here are a few points:

• We are interested in talking to people that are hands-on innovators or bring a novel perspective about data science.

• We prefer to interview people with unique voices that can share opinions not found elsewhere.

• Generally, we like content that leans towards being evergreen. Certainly newsworthy things can be appropriate, but content that holds up a few years later is more interesting to us.

• We try and find guests with good public speaking skills who can be easily understood by our English speaking audience. A review of our archive will show many cases when we violate this idea. We always bias towards good content first and foremost, even if the guest isn't the best public speaker. Yet, this is an important consideration.

• We are only interested in guests that can speak about specific topics related to data science, machine learning, statistics, and artificial intelligence, ideally with a perspective of scientific skepticism.

All in all, this post is not meant to discourage submissions. While we might not be able to reply, we definitely do not want to discourage listeners who are making well intentioned recommendations. Unfortunately, the flood of spammy and mis-leading submissions from PR firms has demanded that we put up a few guidelines.