In my first post (if you read the whole thing, thank you!) it may seem like I painted a very “doom and gloom” picture. The reality is that I’ve never been more bullish for companies and marketing teams. If I thought otherwise I wouldn’t have decided to start AlgoLift. That’s why I’m excited to share in this post more about AlgoLift and what we’ve been building for the past 2+ years. I’d love to hear your thoughts or feedback in the comments.

– Andre Tutundjian, CEO and Co-Founder


Marketing is an Investment

Companies and their marketing teams are in the midst of a transformative crisis: competing interests and rising costs have mandated a shift away from a cost-center approach to marketing. This is not new news: in a 2009 survey, the CMOs of US-based firms said proving the ROI of their marketing was a top priority. That was ten years ago. Today, the need is even stronger: a ROI-mindset is a business imperative.

For more than a decade, marketers and their CFOs have been asking for a return on marketing investment — and aiming to improve it. Naturally, for every dollar a company spends on marketing, there should be an expectation of a positive return.

This mindset — that marketing is no longer a cost center — is clearing the clouded thinking that has prevented smarter marketing investments. Whereas businesses were once satisfied with simply driving impressions, likes, and clicks, it is the inability to effectively measure the business outcomes from these goals that have led to a now widely adopted view that businesses need to demand more from their marketing initiatives. As Harvard Business School professor Thales S. Teixeira puts it, “In general, higher budgets invested in marketing activities with steadily decreasing effectiveness are bound to eventually halt the budget increase.”

This is exactly why we built AlgoLift.

AlgoLift is a financial engine for marketing investments. Our marketing investment platform determines the future lifetime value of customers — from day 0 — and deploys an optimal investment strategy that delivers positive returns. We solve the lack of confidence in LTV modeling and quantify ROI, so marketing spend isn’t the equivalent of backing a dump truck full of coins up to a wishing well and calling it a “cost center.”

Our suite of products is built with a single underlying belief (which I will continue to repeat):

Marketing is an Investment.

The financial industry has made complex investment tools accessible to the masses — Wealthfront, Betterment, Robinhood, Personal Capital — that have transformed the way we approach personal investing. Businesses and their marketing teams deserve something similar.

At AlgoLift, we approach marketing with a financial mindset and our technology operates much like financial technology for marketing (#FinTechforMarTech):

  1. We forecast future revenue (customer LTV)
  2. We forecast future losses (customer churn)
  3. We optimize marketing investment portfolios for ROI (marketing automation)

As business-critical as knowing and maximizing for ROI, it’s not feasible for every company to devote the six months of engineering time (and cost) to developing a model, integrating disparate data sources, testing that model and validating, and then making it actionable. I know because, in the classic founder fashion, I had this exact same problem. Back then, I didn’t have the build or buy question; there was no “buy” option. So, I built instead. That’s AlgoLift.

How AlgoLift Works

Our proprietary behavioral algorithms create unparalleled insights from Day 0 — the first day you see a user, whether they’re immediately a customer or not — that adapts and updates daily.

We developed these algorithms to leverage data that nearly all companies have like user attribution information and transaction history. With just this data (we can always include more), our algorithms build behavioral trends to forecast — as a monetary value — when an individual will make a transaction, the expected number of transactions, the future value of those transactions as well as their unique churn probability. We call this pLTV — predicted lifetime value.

It’s a self-tuning algorithm. Your customer data is analyzed (from any integration source) and we apply relevant algorithms based on your business. Our algorithms forecast key investment indicators: LTV, revenue, and churn. All of this readily accessible on performance and reporting dashboards with tools to create an optimal investment strategy that’s automatically applied.

It’s all re-analyzed daily to continuously improve your marketing investment strategy. The algorithms are extensively validated on historical data to ensure accurate, unbiased estimates of LTV. Since we started AlgoLift our algorithms maintain an average 90% accuracy, 365 days from projection.

When you know how much a customer will be worth in the future, you know how much it’s worth spending to acquire them today.

<21 Day Integration and Daily Customer-Level Insights

As I’ve already written, in-house development of your own model could easily take six months or more; to get insights from that data could take even longer. This also assumes you already have the right team in place to build a working model (not guaranteed) during that period.

Instead, AlgoLift has more than 2+ years proven performance that works with any amount of data (0+ months). We also understand that you’re time and resource constrained, which is why we have a non-intrusive deployment — no SDK, no code — that takes less than 21 days from start to finish.

That breaks down like this:

  1. One day for the project kick-off to determine data availability and establish a timeline
  2. 3 to 10 days for the initial data ingestion, data modeling, and model validation for accuracy and bias.
  3. 3 to 10 days to fully automate the data pipeline and pLTV outputs and grant team access to the recurring user-level data and model accuracy reports.
AlgoLift’s end-to-end integration process

The setup is simple — just how I would have wanted it sitting on the other side: there’s no SDK, no custom integrations, and no code. And the data we collect is also simple (and secure) — with non-PII attribution and transaction history.

Our Algorithm Accommodates Churn and Non-Payer Conversions

Unlike many LTV models, which have difficulty accurately accommodating churn (or customers whose time to future purchase of “never” doesn’t fit the time-based model), we’ve separated our tuning strategies for LTV and churn. This makes our predictions for each much more useful.

Our algorithm also forecasts revenue from future conversions of non-paying users — think future revenue from current Spotify free users. In my experience most companies only forecast LTV for customers that have already made a purchase. This is a big miss. This non-paying (i.e. not yet a customer) cohort represents a significant portion of future revenue — and a significant marketing opportunity.

One Model, Any Business

While there’s just one foundational AlgoLift model, we have built it in such a way that it adapts to the nuances of any business, including those that generate revenue through in-app purchases, ads, subscriptions, in-store purchases, or ecommerce. Our self-improving algorithms forecasts LTV daily and tracks accuracy with user-level LTV and churn forecasts at the most granular level: the individual user (anonymized of course).

Bringing it All Together: The AlgoLift Platform

By now you’ve read a lot about our beliefs, our algorithms, and our approach. All of this manifests itself as The AlgoLift Platform, which is made up of three primary product pillars each associated with core values:

  1. Intelligence Engine: Value of Customers
  2. Optimization Engine: Value of Investments
  3. Data Engine: Value of Trust

Within each of these product pillars we arm companies with tools critical to making marketing investment a reality for their businesses:

Real Customer Examples with Real Business Results

Improve 90-day returns and increase marketing spend

In our experience, AlgoLift makes it possible to simultaneously hit goals that once seemed at odds with one another. One of our largest gaming customers wanted to improve its 90-day marketing investment returns while dramatically increasing marketing spend. This double-sided goal had previously been impossible for them. Either they could scale spend, or they could increase short-term ROAS; they couldn’t do both at the same time.

Our optimization engine combined daily algorithm forecasts of user-level pLTV with delivery algorithms to deliver an optimal marketing investment strategy. The AlgoLift platform reallocated the portfolio of spend across multiple channels and adjusted budgets and bids to deliver real business results:

Minimize CPI, scale spend, and maintain ROAS

In another use case, a mobile-app customer wanted to create and execute a marketing investment strategy to minimize cost per install (CPI) while simultaneously scaling spend to accelerate app growth. They did not, however, want to simply grow with low-quality users, so they also aimed to minimize any adverse effects on 90-day ROAS. AlgoLift delivered on every objective:


In Sum

Obviously, I think the AlgoLift platform and the team making it all happen is truly game-changing. It transforms the idea of marketing as an investment opportunity into a business reality. It solves the problem of rising costs and diminishing returns by solving the LTV and attrition problem with technology. In the age of algorithms and machine learning, it’s a business imperative to use data-analytics to your advantage — so your company and the marketers who drive its growth can stop reacting to outdated trends and stale data and start making investment decisions based on expected future outcomes.

If you have thoughts, comments, feedback, or want to talk more then I’d love to hear from you.