Applying Lean Startup Methodology to B2B Online Advertising Strategy

I’m a big fan of the Lean Startup methodology, and not just for launching a product. Lean Startup methodology includes reducing waste by focusing on rapid testing of minimum viable products, and learning from the results to inform where you should actually invest your time and resources. Let’s bring this back to B2B online advertising to determine which promotional channels will create the most cost-effective qualified leads for your business.

There’s an enormous danger with wholeheartedly embracing Best Practices and believing that those Best Practices are the universal course of action you should take. There is no such thing as “Best Practices” with upper-case. There is only what is your specific best practice for your company, your marketing, your industry, and your database. When someone says that LinkedIn promoted posts are the best way to get leads, and that becomes Best Practice, the effectiveness of promoted post goes down significantly as competition swarms and people become blind to the advertisements. What we need is to be able to test what is most effective for us, not everyone else, at a given moment.

Lean Startup Methodology Guiding your Online Advertising Strategy

In many marketing departments, when people open a new online advertising channel, they invest enormous amounts of resources into perfecting them and squeezing out every optimization they can. It’s 100% investment in a channel because it seems like a good idea. Targeting out the wazoo, a hundred ads, A/B testing custom-designed images, everything. This. Is. Wasteful.

Instead of charging blindly full-force into a promotional channel, we need to stop and think about what we’re doing. There are lots of promotional channels—too many to invest in. Let’s create minimum viable products (MVPs) of each channel, test them against each other, and look at the numbers to see what we could theoretically optimize based on benchmarks and guesswork.

The Process for Rapid-Testing Promotional Channels

What we need is to build a test to determine which channel is better relative to the other channels while controlling as many variables as possible. We gather the operational data (cost per click, landing page conversion rate, percentage of net new leads who are demographically qualified) and run the math to figure out what theoretical cost per qualified leads we could produce if we were to spend time optimizing the channel.

Running the math from your original baseline is the real difference from a normal promotional practice—normally one might just put in the time and energy trying many different things to increase conversion rates, click-through-rates, etc. before doing the math. Instead, we see what cost per qualified lead (CPQL) we would be able to produce, and determine if we should instead spend that energy optimizing channel A or channel B.

So here’s what you do, B2B content marketers:

  1. Pick a gated lead generation asset that performs pretty well for you right now. This is the asset we’ll be testing across channels.
  2. Pick some channels that you’re going to rapid-fire test. Depending on your time/resource availability, you could start with three separate channels. Theoretically, we can try Google Adwords, LinkedIn Promoted Posts, and Twitter Promoted Tweets.
  3. Create your ad copy, keeping each channel as similar as possible while not sacrificing quality. Have your designers design a single piece of creative for each channel—base them off of the same original image, then tweak for size/proportions. You’re battling two forces here: trying to restrict variables while at the same time starting with something you would do effectively for a specific channel.
  4. Target each of your channels as effectively as you can, but you’re running a single campaign right now, promoting one asset. The point is to create a minimum viable product for this ad. The effort for each channel should be minimal, and that’s the point. You’re not scrimping on quality, you’re just purposefully not scaling your setup.
  5. Run your campaigns for one or two weeks, starting at the same time and ending at the same time. Don’t tweak them along the way.
  6. Gather the numbers and see what just happened. Note that these qualified leads are probably going to be exorbitantly expensive, and that’s okay, because we’re just using this to learn which channels we can afford, and which are effective.
  7. Play with the numbers and see what fantasy-land CPQL you could achieve through optimization. This is where you’re saving time and money, compared to the alternative. You don’t have to spend the actual time and mental energy to do the optimization, you just have to know the returns on that optimization before you decide to put the time in, to determine if the time is worthwhile.
  8. Use the results from this test (and the playing with the numbers) to determine an action plan for your online advertising strategy. Prioritize based on CPQL now, in the fantasy-land future, and based on your revenue-model-waterfall’s acceptable CQPL.

Real-World Example and Excel Table for Measuring Online Channel Success

Here’s a pretty realistic picture of what you might see, and the Excel table you need to make:

content-promotion-cpql

Put in your values for CPC, Conversion Rate, and % leads who are demographically qualified. All we care about is the highlighted column—cost per qualified lead. You probably have absurd numbers there, and that’s okay. This is just a pilot. All we’re looking at is relative effectiveness of channels and then what CPQL we might be able to optimize toward.

Start playing with the numbers. What happens if you decrease CPC by two dollars? What happens if you increate LP Conversion rate by a percentage point? How is the CPQL affected if you hone your targeting and increase % Demographically Qualified by ten percentage points?

In my experience, column D (% of leads who are demographically qualified) is the most expensive (and overlooked) conversion. If a channel is bringing in mom-and-pop shops, they’ll never be able to purchase our services, even if they’re interested in the content we’re producing. That’s where, in this case, LinkedIn promoted posts shine, and blow AdWords out of the water.

content-promotion-fantasy-cpql

Playing with the numbers from this test, we have an action plan: Dump money into LinkedIn promoted posts because it’s already an acceptable CPQL without optimization. We can’t afford AdWords even if we were to get heavy optimization gains all around, so it’s not worth our time/resources. Twitter might be affordable if we optimize.

Not all channels are effective for all industries. Not all channels will remain as effective over time. If you do this, you’ll easily see what channels can generate leads cost-effectively. Based on your revenue model’s waterfall, you should have a target average CPQL, and you can use this approach to see which channels get toward that CPQL.

Notes and exceptions (like any true Economics paper): Obviously, there’s more that you can measure here. Average behavioral score per channel is one such example, and can justify higher CPQL in some cases. This is true if your AdWords campaigns are bringing in BOFU-prospects who quickly progress to having positive conversations with Sales—really in that case you’re looking at Cost Per Sales Accepted Lead (CPSAL). If your revenue model is bulletproof as it should be, that will quickly show to be the case. Demographic gating is enormously important in these considerations, though, because a student who has a behavioral score of 1,000 is not representative of marketing’s success.

Pin It
  • http://www.marketingrockstarguides.com/ Josh Hill

    Ed, great article! I had no idea what practical gems were on your blog :)

  • http://grantgrigorian.com Grant

    Brilliant!

    The column D is something that’s very often excluded from this exercise – much props.

    I have a question though – in my experience each of the channel with have diminishing returns to scale. How do you anticipate, or plan for that – if at all? In other words, is there anything I can do to test my “market size”, in the analogy of the lean startup?