Proxima

Proxima

Software Development

New York, NY 1,828 followers

Unlock profitable growth with predictive data intelligence

About us

Helping eCommerce brands scale their business profitably and maximize marketing efficiency. We use machine learning and predictive data intelligence to drive incremental growth and unlock critical business insights. WE'RE FOR BRANDS LOOKING TO SCALE We help brands get the most out of their paid marketing efforts so they can scale ad spend without efficiency loss. Proxima's predictive intelligence tools lower acquisition costs and increase customer LTV. GET STARTED Book a call with our team here: https://www.proxima.ai/demo GOT QUESTIONS? Drop us a message here: [email protected]

Website
https://www.proxima.ai/
Industry
Software Development
Company size
11-50 employees
Headquarters
New York, NY
Type
Privately Held
Specialties
DTC, Ecommerce, Digital Ads, Facebook, Instagram, Data Science, Digital Marketing, Data Intelligence, Advertising, Meta, TikTok, Shopify, AI, Ad Targeting, Consumer Insights, and Benchmarking

Locations

Employees at Proxima

Updates

  • Proxima reposted this

    View profile for Alex Song, graphic

    Building AI to optimize user acquisition at scale.

    Hyped to share our latest case study — this one's a triple threat: Proxima x Astronaut Party Inc. x Simon Pearce 🚀 Despite their efforts to target customers in new regions, Simon Pearce wasn’t reaching audiences with the same purchasing power or intent as the core customer base they've built in the Northeast. But the team at Astronaut Party — a leading eCom growth agency — was up for the challenge. And they knew that solving this problem would require serious data science chops. So, they began experimenting with data-enriched audience targeting. The results? → 31% reduction in CAC → 51% increase in ROAS → 60% in TOF Ad Spend Lukas B. Snelling, President & Founder of Astronaut Party, was initially wary of asking his client to pay for a new tool. So we set them up on a free trial with Proxima to test our AI Audiences against their existing targeting strategy. Spoiler: Proxima was the clear winner. Lukas summed it up perfectly: “For agencies, expanding the reach of your clients’ ads is the easy part. The hard part is knowing which audiences to target; you need sophisticated tools to do this efficiently. Proxima's AI Audiences should be a crucial step in every agency’s playbook.” How’d we approach audience building for Simon Pearce? 1) We ran a comprehensive ad account audit and customer base analysis to understand the core shopping trends of their most loyal customers 2) Our algorithm identified consumers with similar traits and purchasing behaviors across the Proxima Shopper Universe™️ 3) We then built data-enriched audiences that were uploaded to Simon Pearce's ad account via the Meta API (and the underlying customer data refreshed weekly) Experimenting with audience targeting in this fashion opens up a level of customization and precision well beyond the capabilities of Meta's broad algorithm. A strategic approach to targeting presents a serious opportunity for every agency and brand. That’s what we’re developing at Proxima — tech that enables marketers to take back control of their targeting. Read the full case study here ➝ https://lnkd.in/eHpdc_qZ  

    • No alternative text description for this image
  • Proxima reposted this

    View profile for Ashvin Melwani, graphic

    CMO and Co-Founder at Obvi

    We spent $230k+ on our Proxima lookalike audiences last month… Yes, going broad is the go-to strategy. But try this targeting experiment: A) Go broad – let Meta decide who to show your ads to B) Go targeted – upload Proxima’s boosted seed audiences and run LALs Put these up against each other. We found that not only do these audiences scale and compete with broad, but we’re also giving Meta additional data points to find incremental audiences to scale into. The point is – don’t fall into the trap of “best practice bias” There are wins to be had beyond broad. Media buyers should not be ignoring audiences as a tool to leverage in Meta. No matter which avenue you dive into, whether you stick to going broad or experiment with audience targeting, if you test audiences against each other (especially in a CBO) you’re only going to get spend when something works. And we’re seeing a lot of success here. So it’s worth testing something outside of the “go-to” strategy.

    • No alternative text description for this image
  • View organization page for Proxima, graphic

    1,828 followers

    ⚡️⚡️⚡️

    View profile for Alex Song, graphic

    Building AI to optimize user acquisition at scale.

    I’m stoked to announce our newest case study with C4 Energy ⚡️ Spending at scale on Meta, they needed solutions other than just iterative creative testing to keep rising acquisition costs at bay. So, they experimented with scaling through audiences. The results? → 20% reduction in CPA → 20% increase in ROAS → 50% increase in MoM ad spend These are impressive figures, but they’re not unfamiliar to the Proxima team. Will Rivas (Media Buyer at Nutrabolt—umbrella to C4) summed it up perfectly: "It's tempting to just go with ASC, throw in some creative iteration, and let it rip. But when you hit a certain spend level, you tap out some audiences pretty quickly… Audience targeting has become a lost art. Proxima reminded us just how powerful audiences can be.” So, how did we approach audience testing with Will? 1) We conducted a customer base analysis, uncovering the traits and purchase behaviors of C4’s highest-value customers. 2) We then tapped into our dataset of ~65M+ online shoppers and $17B+ in purchases — our algorithm identified similar customers and generated predictive audiences for C4 to target on Meta. 3) We seamlessly uploaded these audiences to C4’s Ads Manager via the Meta API (and the underlying customer data refreshes weekly). The results? A 20% improvement in CPA and ROAS, while scaling ad spend +50%. If you're spending at scale, there's almost certainly going to come a point where your ad efficiency will hit a glass ceiling. Especially if your strategy is rooted in "just going broad" and relying on creative to do your targeting. We're professional breakers of those ceilings. I’m not sharing these numbers in some snake oil effort to say we've unlocked some kind of silver bullet. That’s not what we’re saying. However, this data from C4 reinforces a notion I wholeheartedly believe in: Test audiences like you test creative. Experimenting with audience targeting enables you to scale BEYOND what Meta's broad algorithm can deliver and find higher-intent audience pools that drive revenue. Audience targeting presents a serious opportunity for virtually any brand. What we’re doing all day at Proxima is developing incredible tech to help marketers take back control of their targeting, opening up the door for more experiments that can help your brand get a much-needed edge. Running a DTC brand that's hitting an ad spend ceiling? Let's chat. We offer 30-day free trials. Read the full case study here: https://lnkd.in/eZRGjiMM

    • No alternative text description for this image
  • View organization page for Proxima, graphic

    1,828 followers

    always 👏 be 👏 testing 👏

    View profile for Alex Song, graphic

    Building AI to optimize user acquisition at scale.

    More brands should be testing audiences the same way they test creatives. Here's what I mean. After iOS 14, when audiences on Meta and other platforms became much less effective and going broad became best practice, the line of thinking became: "Creative is the new targeting." There's truth to this. Creative testing and shipping a ton of assets per month should be a priority for brands. But, I think we may have jumped the shark on this. Do you really trust Meta's algorithm to get your ads in front of all of the right people? Or are they just getting you results that are good enough to get you to stick around and keep testing? Their incentive is to get advertisers to spend more. On top of that, creative production is expensive. Whether you're going for high-production value or paying UGC creators to make assets for you, it's easy to run up a bill. If you're going to invest all that, you should give each asset the best chance possible to get in front of all relevant customers. Why should we be okay with burning through assets with a 1/10 hit rate? This is where audience testing comes into play. Sure, test broad campaigns. In many cases, they'll work well. But also test alongside interest stacks, lookalikes, and precision audiences enriched with aggregated first-party data (what we do at Proxima). This approach will get that same creative in front of as many of the right customers as possible, leaving no stone unturned. I want to make one thing clear. It's not audience testing OR creative testing. It's audience testing AND creative testing. The two should happen in tandem. Creative iteration is table stakes, and audience testing is about getting the most out of that creative and scaling more efficiently. What do you think of this take? Happy to answer any questions based on what we're seeing at Proxima.

  • Proxima reposted this

    View profile for Alex Song, graphic

    Building AI to optimize user acquisition at scale.

    What happens when you make the jump from a 7-figure to an 8-figure consumer brand? In my experience: you need to get serious about the less exciting parts of your business. You're evolving into a 'real business.' Growing a brand in the early days is high-octane fun. Sure. It's hard work but with a lot of freedom and few constraints. You get to be scrappy. I've been there with my brands WellPath, Finn, and Grummies. The early days are a blast. But scaling to the next level requires two specific shifts in your strategic thinking as a founder: 1) Prioritization and leveling up of operational systems. This means systemizing operations and growth/marketing by making key hires in those areas (I'll make another post on those key hires at some point). 2) You need to go omnichannel at this stage. DTC alone isn't enough. #1 is straightforward. It’s not easy, but with the right hires, you'll be able to hand off more of the main functions in the business as a founder. #2 is tricky. Selling DTC straight from your website doesn't cut it anymore. You also need to be on Amazon, considering big box retailers (like Target, Walmart, etc.), and thinking about wholesale. Scaling gets a lot more complex, fast. Oh and of course, you also need to keep your DTC machine running consistently with all these moving parts in play. It takes a team and a lot of trust in that team. But when it’s working, it's incredibly fun. I'm curious though. If there are any 8-figure founders & operators reading this, what were the biggest learnings you had when pushing beyond $10M in revenue? PS: I’ll be sharing more about my past experiences running DTC brands, as well as how that parlays into what we’re doing at Proxima—would love to have you following along. 

  • Proxima reposted this

    View profile for Alex Song, graphic

    Building AI to optimize user acquisition at scale.

    Your success as a brand founder isn't going to earn you any extra sleep at night. Here's the reality of scaling past the $10M revenue mark as a consumer brand. As soon as you start to scale the brand, the goalposts move. And as your brand grows, so do your problems. The problems you solve scaling from 0 ➝ 1 create new problems for you to solve to get from 1 ➝ 100. The new problems aren't a bad thing. Founders and early-stage operators are problem solvers by nature. Tackling new challenges should energize you. But if you’re not prepared, it can feel overwhelming. As you scale from 7 to 8 figures, here’s what you can expect: (1) Manufacturing volume requirements and costs will increase. (2) Your supply chain, which has held strong until now, will need a serious upgrade. (3) Post-iOS 14, it's way harder to maintain healthy unit economics if you're heavily reliant on paid ads to fuel growth. Look. Scaling to 8 figures is incredibly exciting. It's a huge milestone you should celebrate. But sustaining it isn't as straightforward as getting there. Keep the above three points in mind and you’ll be on the right track. Question for you: Of the 3 points I mentioned, which would you want to see a more in-depth post on? Let me know and I'll make it happen. PS: If you do want to scale your paid media spend more efficiently, you should check out what we're building at Proxima (shameless plug 😉)

  • View organization page for Proxima, graphic

    1,828 followers

    🌶️🌶️🌶️

    View profile for Alex Song, graphic

    Building AI to optimize user acquisition at scale.

    This might be a hot take, but my position is that the obsession with mass creative testing in ecommerce is misguided. So you're telling me...we're supposed to just keep throwing asset after asset into an ad account and keep our fingers crossed, hoping for the best? Ad platforms like Meta or TikTok are a black box. Nailing 1 asset out of 10 tries is deemed a 'success.' And to me, running a creative strategy this way feels wasteful. More and more, marketers and creative strategists are being tasked with jumping on a creative treadmill and never coming off of it. I don't know. Maybe I'm off base here...but curious to hear what you think. What's your take on the state of creative testing for DTC brands? PS: I'll be sharing more of my thoughts on this, and how we think about scaling ad accounts at Proxima in future posts.

    • No alternative text description for this image
  • View organization page for Proxima, graphic

    1,828 followers

    *taking notes*

    View profile for Ashvin Melwani, graphic

    CMO and Co-Founder at Obvi

    I’ve seen so many people time and time again not squeezing allll the juice out of Meta Listen… There are a lot of nuances to scaling ‘efficiently’ – especially once you cross the 6-figure mark in monthly ad spend. But to maximize the value you get from Meta, you need to prioritize 3 core pillars: 1️⃣ Campaign Setup 2️⃣ Audiences 3️⃣ Creatives There’s no one-size-fits-all playbook, but these pillars are a constant area of focus for the best media buyers I know. I’ve talked a lot about account structure and creative iteration, so today I want to start digging into audiences. 9/10 marketers will say “just go broad.” They’re not wrong. But I want to reframe the conversation on audience targeting. It’s not just this OR that. It’s this AND that. Go broad AND experiment with other audience types in Meta. Efficient scale is the name of the game – so why would you handicap yourself from testing something that could unlock that? It also doesn’t hurt to diversify your targeting strategy. Something we’ve been experimenting with lately: AI-powered lookalikes. Instead of building seeds off of our customer lists or using platform engagement data, we’re flooding the zone with boosted seeds built from a ridiculous amount of aggregated storefront data. The volume and fidelity of this data is what’s enabling us to use lookalikes to target high-paying shoppers that are in-market with HIGHLY targeted ads. This gives us a much more surgical, targeted approach with a lot less wasted ad spend. To date, we’ve seen a 31% higher NC-ROAS and 26% lower NC-CPAs using lookalike audiences compared to just going broad, and I’m at a point now where I’m allocating a little over 60% of my total ad spend to these audiences because of it. Listen, if you’re spending more than $100k a month on ads this is something I highly recommend. There are a few tools out there that can help you build these types of audiences, but I personally recommend Proxima Their audiences continue to outperform and their team is second to none. They build audiences with custom AI models and data from over 65 MILLION high AOV shoppers (and over $17 Billion in transaction data) across 1000s of Shopify stores. Yes, I’m a #proudpartner and some will say that I am biased, but the results I just shared? Not a fluke. I wouldn’t be on here promoting them if it was. They have helped us consistently scale our ad spend more EFFICIENTLY with more PROFIT overall, so of course I’m going to share it. Wouldn’t you? PS: If anyone wants to see actual screenshots from the ad account of all numbers mentioned here - feel free to DM.

  • Proxima reposted this

    View profile for Alex Song, graphic

    Building AI to optimize user acquisition at scale.

    I’ve been a successful brand founder, launching and exiting an 8-figure supplement brand. I've also been an unsuccessful brand founder, shuttering one of my brands ~1.5 years in. Like most founders, I’m familiar with wearing both hats (heads up—neither hat is necessarily comfortable). I’ve done a lot of learning as an operator—probably more learning than coaching. Some of this learning is fun, some of it is painful. But these experiences have been invaluable. A couple of the biggest learnings I’ve uncovered (and what led to the creation of Proxima in the first place) are 👇 → In a post-iOS 14.5 world, traditional DTC CAC/ROAS models stopped working.  → Customer acquisition became more expensive (too expensive for many) and it legitimately led to an adapt-or-die environment. You NEED to be ready to adapt. And I know. You’ve heard it before. But trust me, it’s better to “hear” that sentiment than to experience it—at least the “die” part. I built Proxima with the above learnings top of mind. To lower CAC and increase LTV through sophisticated predictive data intelligence—regardless of market fluctuations. It’s led to our recent Series A, but we’re only just getting started. I'll be sharing more of the learnings during my time running consumer brands and more of how we're thinking about scaling Proxima post-Series A (hiring, tech updates, etc.). I plan on getting a lot more into the weeds of what it really takes to run a successful brand in 2024, and how to use data to do so. If that sounds compelling, would love for you to follow along.

    • No alternative text description for this image

Similar pages

Browse jobs

Funding

Proxima 1 total round

Last Round

Series A

US$ 12.0M

See more info on crunchbase