There are several ultimate guides available for running ads, and I have no ambitions to create another one. So here’s a penultimate one, containing only six pieces of advice. But the advice is practical, to-the-point and based on more than 10 years and several million dollars of ad investments, as well as several million dollars of ad spend (yes, these are different things).
1. Make the first pick of channels, using category awareness and category urgency as your guide
“What channels should we use?” is a pretty common question and while all businesses are different and there is no best answer, I’ve found it helpful to use category awareness to make the first pick of channels.
Category awareness is high when most people in your target audience already know the product or service you’re selling. It’s things like car insurance, CRM software, web hosting, smartphones, etc. Each of these things is generally well known and people in the target audience use search engines and specialist shopping or comparison sites when they feel a need.
On the other end of the spectrum, there are things that most people don’t yet know to exists. Things like heart-rate monitoring jewelry, applications of a brand new technology (eg “serverless” in 2015) or products like Airtable which are difficult to put into any category. If no-one knows the product category, then no-one is searching for solutions in that category. (I’m pretty confident that there are no searches for “software that is a little bit like a spreadsheet and a little bit of a workflow tool and a little bit of a database.”)
In addition to category awareness, there is also category urgency which indicates the level of active buying in the industry. For example, awareness of DVD player category is high, but urgency is rather low. And for gym memberships, awareness is high throughout the year, but urgency peaks around the first two weeks of the year.
If category awareness and urgency are high, people search. If people search then, wait for it, it’s a good idea to use search and ads on sites that come up in search.
If category awareness and urgency are low, display and social ads that target based on proxies like interests or demographics can drive results.
It’s worth pointing out that many products are hybrid, some people in the target audience know the category well, others need to be educated. You’ll then need to cover both types of channels, but you’ll likely find search and high-awareness channels to be more profitable and/or faster to drive results.
2. Ignore price-per-click and use profitability metrics instead
“Does it make sense to use channel X if price-per-click is $10?” is another popular question. This question doesn’t compute more than asking whether green is a good colour. Price per click is meaningless. What you’ll want to know instead is CPA (cost-per-acquisition) and its ratio to life-time-value.
Here’s an example:
- Let’s say we’re marketing a hypothetical email product that costs $100 per month, users usually stay on for 20 months, so LTV is $2000
- We’re using Adwords to advertise, and we’re only using two keywords: “cool products” and “email tool for ecommerce”. Clicks cost $2 for the former and $10 for keyword “email tool for ecommerce”. We picked the first keyword because it’s cheap and kind of relevant. (Because we have a cool product, right? Right?)
- Clicks convert to free trials at 1% rate and 5% rate; free trials convert to paying customers at 5% rate and 15% rate.
- Let’s keep things simple and assume LTV is the same for both keywords (which rarely happens in real life).
While price-per-click for “cool products” seems more attractive, you’ll spend more than you earn if you put your budget behind that keyword.
You’ll need real data to be able to make this calculation. Which leads me to my next suggestion. You’ll want to:
3. Start paid marketing as early as possible
At Pipedrive we started experimenting with a 500€-1000€ per month Adwords budget as early as there was some spare change available. This drove some signups and got us more customers but more importantly, this gave us data about what worked and what didn’t.
Despite having to turn off all paid advertising during one or two cash-strapped patches, by the time we had secured first rounds of funding we had enough data to conclude that some channels or keywords groups returned money in less than 6 months. We also had learned that others that seemed equally attractive on cost-per-signup basis performed 2-3 times worse due to lower conversion to paid, lower revenue-per-customer (ARPU) or higher churn.
In the example above, some channels and keywords seemed very promising or extremely unprofitable based on the first couple of conversions. We arbitrarily set a limit of 25 conversions per keyword before we concluded something as a win (or a loss).
Depending on the industry and available budget it takes 6-18 months to get meaningful data about churn and ARPU of a channel or ad group. You’ll want to have that data when you have the budget to scale which means you’ll want to start experimenting many months before you have a budget.
And this is a good segue into my next recommendation.
4. Be comfortable in wasting 10-50% of your budget.
“I know that half the money I spend on advertising is wasted. My only problem is that I don’t know which half.” – William Hesketh Lever, founder of Unilever.
Using online channels and basic tracking you’ll still end up “wasting” some of your advertising budgets, but you’ll know rather accurately whether something worked or not. (I’m excluding the broader topic of brand advertising here and only talking about direct response ads). And depending on your budget and level of aggressiveness you can choose the level of your “wasted budget”.
The “wasted budget” is not really wasted, because it tells you what creatives, channels and keywords are not working. It’s more your education fund than your marketing budget line.
The more comfortable you are to “waste” money, the faster you’ll learn. If you have high margins or sizeable coffers of funds, you can choose to temporarily or permanently put as much as half of your budget into learning, but I can’t think of many situations where one would want to spend less than 10% of marketing budget on experimentation and discovering new channels.
5. Set your advertising budget based on payback time and runway
Another popular question: how much should we budget for paid ads?
Here you need to know two things: what’s the average payback time of your ads and how many months of cash is left in the bank. As long as the latter is bigger than the former, you should spend as much as you can (without sacrificing efficiency of your campaigns).
If however, payback time is longer than your runway, you’ll want to be more conservative. And your level of conservatism depends on how close you are to running out of money as well as how the payback to runway ratio has been trending.
Please note that it’s possible to run highly profitable ads that you can’t afford. For example, if your campaigns have an ROI of 400% and this return manifests itself over 3 years, but you only have 9 months of money left, you should cut all channels that return money slower than in 9 months.
6. Use agencies when and where it makes sense
Finally, the question: should we hire in-house or hire an agency? It depends.
My rule of thumb is to manage core channels in-house and outsource experiments to an agency. At Pipedrive we had a significant paid budget. Two channels accounted for 60-80% of our paid spend and 80-95% of our results. One of these two channels was Adwords which was managed both in-house and by an agency and there was a clear difference in results and profitability when they were managed in-house.
On the other hand, if you’d like to follow tip #4 above and invest in learning new channels (think new ad formats on social platforms that pop on almost every week), it’s simply inefficient to spend time on learning all of them. You’re much better off handing these off to an agency that already knows that format or platform. Then, if the initial tests reach their CAC or LTV goal, you can declare that channel as “core” and bring it in-house for further optimization and scaling.
A shout out to Bizmut who have helped me scale testing for several channels. They’re one of the very few agencies I’ve worked with that know the SaaS world inside out.
I hope to read your feedback or comments on GrowthHackers, or below.
PS. If you’re driving paid leads and happen to be using Pipedrive, check out Outfunnel, my new startup.