Growth Engineering - Ladder's secret weapon

Ladder’s Secret Weapon: Growth Engineering

Since 2014 Ladder has worked with over 200 companies to help them grow — from Fortune 500 companies to fast-growing startups. Here is the explanation of what Growth Engineering is how it has helped us succeed.

Like in any competitive activity, mastering the fundamentals is key: for growth that means knowing your marketing funnel metrics inside out, memorizing a list of tactics that work for each industry, and understanding your audience so that you can craft a meaningful strategy.

Having all this is important for any growth strategist to stay in the game, but it isn’t always enough — sometimes you need an edge, a secret weapon. For Ladder that has been growth engineering.

What is Growth Engineering?

Growth hacking was traditionally defined as what you get when you mix a developer and a marketer. It was a term coined over a mint julep by Sean Ellis, to differentiate what you got with a traditional marketer (TV ads, billboards, branding) from what he and his contemporaries in Silicon Valley were offering (conversion optimization, digital ads, and viral loops).


For Ladder that meant investing in a technology team as soon as we were able to afford one, and expanding into analytics, tracking implementation, landing page building, campaign automation and data science as we grew. We built a team of smart people and led innovative experiments into ways to scale our client’s growth.

For many years our growth engineering work was our secret weapon: clients would tell us they chose Ladder as a growth agency because we had built our own tools, and we were able to keep many key clients by building custom solutions for their growth problems. Most marketing teams even at fast-growing companies don’t have access to developers and having even one on your side can be a huge competitive advantage.

In this post, I wanted to walk you through the history of Ladder’s growth engineering efforts and explain what we did, as well as what results each major project achieved. We have had to anonymize these case studies so as to not share sensitive information, but you’re here to learn what growth engineering can do for you, and get ideas for what to experiment with.

Growth Engineering Case Studies

Experiment Management Platform for Ladder

The first software we built was for ourselves — we were following Bryan Balfour’s Growth Machine process, but that left us with hundreds, approaching thousands of experiment documents in Google Docs. They weren’t easily searchable and as the team expanded inconsistencies in formatting and fields crept in.

Ladder Planner

So we built our own ‘Trello for marketing’ style tool, the Ladder Planner, which would enforce a specific standard and store all test documents in a searchable database. This was the bedrock on which we built our agency, and the main thing that built trust with our early clients. This also powered the famous Ladder Playbook — having all our experiments in one place, as structured data, made it easy to publish new ideas to the database.

E-Commerce Tracking for a Fortune 1000 Fast Food Chain

Though we were building technology for ourselves, we rarely in the early days did anything technological for our clients — outside of using 3rd-party tools like Zapier. The exception came when a client needed a tracking and analytics setup, which we found ourselves doing late one night for a major fast food chain.

It was the day before a big launch and the agency in charge of building the website came to us in a panic — they had nobody in-house who knew about Google Tag Manager or setting up tracking, yet the client had asked for it last minute. This is where having growth engineering chops can help: we quickly spun up a full advanced e-commerce setup for Google Analytics, tested it was working and published it to production just before midnight.

Annual Budgeting for a Billion-Dollar Public AdTech Company

The second area where an engineering mindset can help is in forecasting future performance. Most agencies simply accept the goals handed down to them from the client, or do a simple extrapolation of the current growth rate and hope it holds.

We wanted to be more data driven than that, and our client was a billion-dollar adtech firm, so we built a full blown monte-carlo simulation to work backwards from their goals to what their recommended budget should be ($1.2m). The benefit of this method is you get a range of probabilities of what goals you’re likely to hit given different budget levels, and can choose how much risk you’re comfortable with. We did this manually in Excel, but now we use a no-code tool called Causal to do it for us (and make better looking visualizations).

Automated Campaigns for a Global Events & Attractions Publisher

Where we really graduated into growth engineering is when we started to build tools to automate some of the repetitive work we were doing for clients. We kept these activities separate from the product team (who by now had external customers to serve) and hired our first developers.

One such project that kept coming up was automating the creation of hundreds of campaigns on the Google and Facebook ads platforms. For a global events publisher we had a feed of incoming products (event tickets and deals on attractions) and used that feed to generate ads from a template. The system would automatically geo-target to the right location, and turn off campaigns if the client ran out of inventory.

Dynamic Landing Pages for a Startup Trade Services Marketplace

Once you automate campaigns, you start to realize the need to automate landing pages too — when you have thousands of keywords it’s important to match every keyword to a unique page. This simply isn’t possible for most teams, so we would call in our Growth Engineers to help.

One such project was for a trade services startup — using Google Optimize we wrote JavaScript that dynamically changed the headline and hero copy to match what was in the ad, based on 17 different pre-programmed categories. it would read the campaign name from the utm parameters in the URL, then serve the page copy that fit that bucket. In many cases this would result in double digit improvements in conversion rate.

Cohort Analysis for Mobile Banking App

After a while just using Google Analytics doesn’t cut it — we found many clients needing more advanced reporting capabilities as they scaled. For example for one fast-growing fintech client, we were spending 5 figures in budget per day, yet it would take in some cases weeks for customers to complete the signup flow.

The solution was building a custom cohort analysis report. This showed us that although any one day was noisy in the data, when you look over time, the number of conversions you get on day one was highly predictive of the number of conversions you’d get in the future — as much as 95% accuracy! So going forward we were able to optimize that client’s campaigns based on 1 days worth of data, and check that custom report to investigate the true underlying patterns.

Custom Reporting Platform for Ladder

With more and more clients needing custom reporting that wasn’t possible with off-the-shelf tools, we expanded from the Ladder Planner to a tool we called Spotlight. We built connections into the major ad platforms as well as a few organic sources and audit tools, and let users drag and drop to build their own reports.

Ladder Spotlight

This was essentially an early Google Data Studio, yet before that was an option and BI tools cost tens of thousands of dollars a year. At one point we had over 60 client reports being automatically generated every Monday, pulling from 10+ sources using proven templates we knew would show the right data at the right time.

Custom Attribution Model for a Fashion Affiliate Startup (Acquired)

With custom reporting often comes custom attribution — once you can see your data in any way you like, you tend to realize you don’t have the data you wanted! So for many clients, we ended up building custom attribution systems. This was possible through custom first-party cookies and hidden form fields, however now server-side tracking is opening this whole field up.

In the case of one fashion affiliate startup, as much as 40% of conversions actually had a touch by Facebook ads, whereas previously they were being misreported as direct on a last-click model. This was game-changing for the client which increased their investment in Facebook ads and were well on their way to a big-name acquisition.

ML Creative Tagging for a Fortune 500 Travel Company

Remember when we mentioned we built a tool just like Google Data Studio? And before that, we built a tool we described as “Trello for Marketers”? Well, lesson learned — both of these products eventually dropped out of usage, first externally and then even internally, as we didn’t have the resources to keep up with the competition.

We were still very happy with our software investment because at the time it was our edge in beating other growth agencies to the pitch, but it came to the point where our custom software started losing us deals because they preferred to use their own productivity and reporting software. We decided to make a hard pivot and switch off these first two products, in favor of doubling down on one obscure feature we had built into Spotlight — ML creative tagging.

The way it worked was by pulling all of your ad creative from Facebook, then using machine learning to annotate what was in the images, you could get performance by tag. That allowed you to see that, for example, dogs performed better than cats, pizza better than Chinese food or landscapes better than illustrations.

For one Fortune 500 travel brand, we found that ‘office furniture’ performed 66% better than ‘suitcase’ — of course if you’re booking business travel you don’t want to see a bunch of ads with people carrying suitcases through airports! You want to see pictures of people like you, the office managers booking travel from your desk. This was the type of analysis we’d never have been able to do manually because for this client we were running 100s of ad creatives per month — we would have never spotted the pattern.

Growth Engineering at Ladder Today

You might be shocked to know that Ladder no longer maintains its growth engineering efforts in house! It’s true that growth engineering in some way got us to where we are today, but as we scaled our business, we began to realize that some activities were better suited to specialist partners. Where Ladder’s core strength lies, is in three activities:

  1. Strategy — mapping out your marketing funnel and channel mix and measuring against benchmarks to understand where the opportunities are for growth.
  2. Creative — understanding your audience personas and value proposition, then crafting multiple unique ad formats and variations for rapid iteration and testing.
  3. Operations — getting campaigns live from end to end and optimizing them based on strategy and experience in the channel takes a lot of high-skilled, hard to automate work.

These are three skill sets that are universally needed across companies looking to grow with Ladder. Whereas growth engineering is not — marketing teams get the most out of investing in technology when they’re in the scale phase of growth. You’ll recognize you’re in this phase because you’ve picked all the low hanging fruit and are now hitting diminishing marginal returns. As performance starts to decline when you ramp up investment in a channel, then it’s time.

That means you should run through the tactics or strategy manually for as long as possible, with a partner like Ladder or with your in-house team. It often takes a significant investment to get technology to the point where it comes close to human efficiency, and you can take the risk out by doing it a few times manually first. Once you know exactly what strategy works, only then does it make sense to automate.

So how does Ladder handle it’s Growth Engineering tasks now? When a function is highly specialized and isn’t universally required across clients, that’s the perfect candidate for teaming up with an expert partner! In this case, however the partner is a little closer to home — Saxifrage was founded by the former COO & Co-Founder at Ladder, Michael Taylor.

Having led Ladder’s product, data science and automation teams, he’s now working full time on our growth engineering projects, both for clients and internal tools, including the ML creative tagging tool. This means we can still continue to invest in opportunities for innovation as they arise, but we can do so with a trusted partner that is focused on excellence in this domain.