Founded in 2016 by José Vicente Aguerrevere and Max Pedró, Takeoff Technologies popularized the concept of automated micro-fulfillment centers (MFCs).

MFCs were originally touted as a cost-effective solution for grocery fulfillment. Grocers spend significant amounts on store labor to fulfill online orders manually, and the use of an automated system in-store has the potential to significantly cut costs. For example, a human picker could pick ~150 items per hour, while an MFC could fulfill 600 items per hour. In theory, you could reduce labor costs by 75%, which, in many cases, could justify the upfront cost of the solution.

However, fast forward to May 2024 and Takeoff Technologies filed for Chapter 11 bankruptcy protection amid a decline in demand for grocery fulfillment automation solutions, which was particularly true for micro-fulfillment centers.

In this insight, we’ll be exploring what went wrong. We’ll then dive into how the market is evolving and developing, and, finally, we’ll discuss how solution providers can better position themselves, learning from the mistakes of the first-generation automated MFCs.

What went wrong?

Several factors led to the negative perception associated with automated MFCs. One of the main reasons was the decline in the online grocery market following the end of the Covid-19 pandemic. However, we also saw that the ROI wasn’t as attractive as many grocers originally thought.

The decline in online grocery sales in the wake of the Covid-19 pandemic contributed to the slowdown in MFC deployments. For example, in May 2021 US consumers spent $7 billion on online grocery, but this fell to $6.8 billion as of May 2024, according to Brick Meets Click. This was driven partly by shoppers returning to brick-and-mortar stores, combined with the cost-of-living crisis. As a result, grocers spent less on fulfillment and focused their capital spending on upstream store replenishment activities.

In addition to the decline in grocery sales, retailers also found that the ROI for automated MFCs was less attractive than initially thought. There were several productivity inhibitors that reduced overall fulfillment throughput – as well as many unexpected costs – which led to the solutions being more expensive than originally believed. This led many grocers to lose faith in solutions and sentiment towards automated MFCs hit rock-bottom in 2022 and 2023.

Productivity inhibitors included inefficiencies surrounding the replenishment process and a lack of coordination regarding the consolidation of manually picked items and items picked from the automated MFC, resulting in lower system-wide throughput rates. The unexpected costs included higher construction and planning costs, longer-than-expected install times, and, in some cases, the misplacement of inventory. The combined effect of productivity inhibitors and unexpected costs significantly reduced the ROI for the first-generation MFCs.

However, whilst the PR associated with automated MFCs is down, we’ve actually not seen a significant collapse in the number of deployments. In fact, as the chart shows, we actually saw the market grow in 2023. What’s changed since the peak in 2021 is the types of MFCs being deployed, as we’ll discuss in the next section.

How the market is developing

While sentiment towards MFCs is at an all-time low, we’re seeing some positive developments in the industry. One of the key trends we’re observing is the right-sizing of MFCs. The first iterations of MFCs were ~30k – 50k sq ft facilities bolted onto an existing store, which serviced both convenience and weekly-shop orders. However, we’re starting to see a bifurcation in the way grocers handle weekly-shop and convenience orders from a fulfillment standpoint.

On the one hand, automated MFCs for weekly-shop orders are getting bigger. For example, H-E-B has shifted away from its smaller 50k sq ft facilities and is instead deploying larger 100k sq ft facilities that can service multiple stores. Weekly-shop orders require a larger assortment of items and there’s often less urgency around fulfilling orders, allowing grocers to centralize operations more, which lowers the cost-per-pick.

At the same time, grocers are coming up with new models for convenience orders. For example, Tesco, a UK-based grocer, has launched a new grocery model called Whoosh. It is a limited assortment platform for online convenience orders, with items picked in its small-format stores located ~1 mile away from the customer. Tesco’s weekly-shop orders, on the other hand, are fulfilled in its larger superstores (many of which now include an automated MFC) with much larger delivery radiuses. By separating convenience and weekly-shop orders, Tesco has created two separate fulfillment models and the automation requirements for each model are very different.

In addition to the bifurcation of convenience and weekly-shop fulfillment models, we’re also seeing more interest from grocers in deploying center-store automation solutions. Cuhaci Peterson, for example, is working with several grocers on architectural designs to automate the center-store. Furthermore, as part of our research, several of the top 10 grocers in the US confirmed that this was in their technology roadmaps. The idea is that in-store customers would key the items they want from the center store into a computer monitor (or place the order before arriving), and then spend more time in high-margin areas of the store such as deli counters, pharmacists, cosmetic counters, or fresh produce.

The right-sizing of MFCs, coupled with the use of more central store automation solutions, is expected to have a positive impact on the market. While we’re certainly less bullish than we once were towards automated MFCs, we’re still confident that there will be growth in this market over time.

How to Effectively Position MFC Solutions

Over the past 6 months we have been interviewing grocers and solution providers to identify characteristics that will enable true economic value to be created for grocers. The insights gathered from these discussions are outlined below with suggestions as to how solution providers can position themselves for success.

First, solution providers need to solve the complexities of online grocery fulfillment with software and augment the solution with hardware where necessary. Too many vendors do the opposite by trying to solve the problem with hardware and then add a thin software layer to support the hardware. A flexible and sophisticated software stack that accounts for the complexities of online grocery needs to be the foundation of an online grocery fulfillment platform. Hardware should be viewed as an add-on where throughput rates are particularly high.

Second, vendors can’t ignore manual in-store fulfillment. Manual in-store fulfillment software needs to be a central part of the solution, and the consolidation of manually picked items and items picked using automation need to be seamless. Right now, this process is disjointed and inefficient, resulting in the system-wide fulfillment rates being far lower than initially thought. For example, one grocer we spoke with installed an MFC with a pick-rate of 600 picks per hour. However, because the manual in-store pick rate was <100 picks per hour, and the consolidation of manual and automation picked items wasn’t optimized, the system-wide pick-rate (across both automation and manual systems) was just 250 picks per hour.

Third, solution providers can’t ignore upstream and downstream workflows. MFC replenishment, inventory placement, and order delivery routing are critical factors that influence the ROI of the solution. Even if the solution doesn’t have these capabilities built in-house, an understanding and visibility of these processes is vital to ensure a robust ROI.

Final Thoughts

Whilst the perceptions of MFCs are at an all time low, the number of deployments are higher than most realize, with 39 deployments in 2023, up from 33 the year before. Furthermore, the opportunity for MFCs is growing substantially over time due to the rising demand for shorter delivery times. In 2024 alone, it’s expected that grocers will spend $18.5 billion on store labor for online order picking. By comparison, the total warehousing sector will spend ~$180 billion in labor, meaning that grocery in-store fulfillment labor cost equates to 10% of the total warehouse labor costs.

Furthermore, the unit-economics remain attractive in the absence of productivity inhibitors and unexpected costs. Assuming that MFC vendors address the key barriers to adoption (which is the baseline assumption in our forecasts), we anticipate more than 1,000 automated MFCs will be deployed during 2030, up from a peak of 58 in 2021.

However, for the market to truly take off, the barriers to adoption need to be addressed. We’ve spent the last 6 months interviewing retailers, automation vendors and solution providers to understand what these barriers to adoption are, and how they can be practically solved. To learn more about the report, click here.

For more information on the MFC market contact Rowan Stott, Research Analyst at Interact Analysis.

*Assumes that 8% of an order value comes from picking, and assumes that 90% of online orders were fulfilled manually in-store. Market size for online grocery in the US is expected to be $257 billion in 2024 ($257 billion * 8% * 90% = $18.5 billion)