Happy Friday!
Yesterday, I tweeted before recording a podcast with Moiz about what we should talk about, and a few questions were about TV… so I figured let’s just write a Friday email on it, right?
In 2017, the butteriest bread and butter was running lower-funnel direct response advertising with the 6-second quick cuts — testing hook after hook, to see who will watch more than 3 seconds, and eventually click through to the PDP. At one point, we started running ads so well, that we got our account shut down. We were getting a 13% click-through rate from Facebook to this article.
Right then and there, we realized, “If we don’t diversify our media mix, we can end up completely f’d!” We were spending north of 7-figures per month on a single channel, with the second largest channel being about 10% of that. Keep in mind, we grew to 7-figures quick… it wasn’t planned to be spending that much. I was just a really good media buyer and we went from $5k per day to some days spending $120k per day. Immediately, we were asked by the executive team to figure out how to make TV work for us.
We knew from spending 7-figures per month what worked — it was the founder story. In fact, nothing came close to the way the founder story performed for us. Couple that with ads being whitelisted from the founder’s own social handles, and we were on fire. So, how could we do that with TV? There were a few key pieces:
Concept & story
Most TV commercials are plain and boring. They have a pretty generic story line, they feel cool to the brand marketing team, and they sell you the product at the end of something they hope will have you hooked in. The reality is — that worked 40 years ago. Today, your commercial has to be something of entertainment value from the beginning.
For us, we looked at what was doing well on every other channel, which was the founder story, and then we decided to double down on that. It worked flawlessly (and continues to work), so let’s go with something we are confident will perform well.
This Hexclad ad is a perfect example of what a great TV ad looks like — entertaining from the first second. And when you get the advantage of TV viewers watching 15 seconds, versus on Meta trying to get past the first 3 seconds, it makes it much easier to convert.
Production
The actual production of TV ads used to be the barrier to entry to get a test going. If you look at the Hexclad ad, that’s obviously a high-production value ad. But if you look at this MVMT ad, you can tell it was produced on a tighter budget. The animations and renders help a lot with increasing the perceived value of the production, and the main filmed portions could’ve been shot with relatively low overhead.
The first TV commercial I was a part of was about $100k in production costs. The third one was $10k all in. It really comes down to how well you architect the storyboard, where you have CGI (computer-generate imagery), and how good your editors are.
You can see the first Hint commercial here, let me know if you find me at 35 seconds!
Media Buying
Media buying is where you take your commercial once the final version is done. At this point you want to make sure that not only is the team buying your ads, or the platform you’re buying with, a legitimate platform, but you should also have the rest of the funnel setup.
In addition to using platforms, you can also look into buying ads directly with outlets like First Media, you can test directly into their channels and see what works. With JUDY, we learned that the Baby First TV channel was the best performing channel, and we could always rely on lower CPAs than Meta, but their other channels didn’t perform for us.
TV reminds me a lot of the native networks like Taboola or Outbrain. There are channels that will work well for you (like on Native, there are certain sites that do well for your brand) and there are those that won’t, and you shouldn’t spend extra dollars there.
Site experience
The commercial and the placement of the commercial is one aspect. But you have to recoginze the rest of the funnel as well — the website, the coupon code, the attribution, the customer service, what does the content look like for someone coming all hyped up from a commercial, etc.
My preference is in the ad to air channel-specific versions or time-specific versions with unique coupon codes so you can see the attribution in real time. If you used JUDY (dot) co / {keyword here}, then you can track the keyword back to a time when the ad aired, the landing page it went to, the bounce rate of that user, the conversion rate from that ad, the customer cohort from a specific ad, etc. The same goes for coupon codes, although those can only be tracked for people who bought, vs just visited.
My thinking is that if you’re running TV traffic, you should focus on holding someone’s hand down the funnel to purchase.
With JUDY, we did something where you would go to the URL: JUDY dot co slash TV. That URL would then redirect you based on the geo of your IP address. So if you entered that in California vs New York, you’d go to 2 different landing pages. In JUDY’s context, we did that because you would have different natural disasters that you wanted to get information for — we were pushing emergency guides you could grab, based on your location (click the states above to see examples).
Through this, not only did we collect a ton of information on the traffic and which channel of airing worked best for us, but we could capture more people at the upper funnel level, versus just catering to those who were going to buy right away.
TL;DR: As much as your commercial is important, you need to have a proper journey that viewers go through, to get to the other side of the purchase.
Attribution
Most big brands (Bacardi, P&G, etc) they run ads, and then run the analysis of what worked and what didn’t — mainly because they don’t get realtime data. In this case, you DO! There’s no perfect way to look at TV attribution, but you have a few data sources to look at and understand what is working and what isn’t working:
- Publisher-direct data: This is feedback or data you get from the channel, if you buy direct.
- Attribution platforms: If you use something like Northbeam, you can run proper media-mix modeling with TV ads, and how that affects numbers on Meta or Google.
-
Buying platform: If you run ads with a platform like Tatari, you get a good amount of first-party data in there to look at and understand how ads drove traffic to your site.
- Old-fashioned way: Using Google Analytics and surveys (“How did you hear about us?” at checkout).
Attribution isn’t a perfect game, there will always be a small gap — but the data should be looked at as data for you to make a decision, not just the source of truth.
Everything I’ve stated above has rung true for tons of DTC brands for the last few years, and still is true today. Now, let me tell you about the best digital-first platform to buy these TV ads.
Tatari — The only TV ads platform that allows you to buy linear and streaming TV, but also doubles as a TV attribution platform using metrics comparable to digital (CPV, CAC, ROAS).
Running TV ads is not super simple. You have the ad tech, creative, attribution, multiple buying methods for streaming vs linear, and generally, most ad platforms aren’t optimized for performance marketing. This is where Tatari comes in.
With Tatari, their platform makes it easy to bring TV advertising in-house. You likely already run campaigns on digital platforms like Facebook and Google, and now using Tatari, you can test and learn across TV in a similar fashion, using real-time data and feedback loops to optimize for the lowest CPA.
If you’re spending money on paid social and search, after a certain point, high CPMs and saturation lead to diminishing returns. That’s where TV’s ~$3 CPM gets really effective, especially when reaching new audiences that make your lower funnel tactics more efficient.
Whether you’re ready to start buying TV ads from the only Shopify Plus-approved TV partner, or you want to still learn more and see if TV is right for your brand for 2023, click here to learn more about Tatari & schedule a demo!