How Fast Is Altera? Benchmarks for Large Imports, Updates, and Exports

A lot of stores ask if Altera can handle their catalog. It can, and the rest of this post is the numbers behind that.

We benchmarked the operations stores run most often. That’s order exports, product exports, bulk variant updates, and first-time imports of orders and products. Each test ran multiple times across different days, so what’s below is medians with min/max ranges and not the best run we got. Headline numbers first, then the breakdown.

The headline numbers

#

For a typical Shopify store, the operations you run most often are the fastest:

• Export 1,000 orders: 38s
• Export 1,000 products: 59s
• Update price and compare-at across 2,111 variants: 3m 18s

The bigger one-time jobs (migrations, historical loads) take longer but still finish in under half an hour:

• Import 1,000 historical orders: 15m 31s
• Create 1,000 new products from scratch with full image processing: 27m 10s

For comparison, hand-editing 1,000 products through Shopify admin (open product, edit, save, next) is a multi-day task. The realistic alternative to a bulk tool isn’t a faster bulk tool, it’s losing days of someone’s time.

The rest of this post breaks each of those numbers down with full ranges across multiple test runs, what shaped them, and what they’ll look like on your own store.

Methodology

#

We ran each test multiple times across different days and times of day. We report the median across runs along with the min and max so you can see the spread.

All tests ran on a paying Shopify Basic store with Altera on the Pro plan. Each run started from a clean slate, no warm caches or carryover state, and we spread runs across different times of day because Shopify’s API performance varies a lot through the day. The min/max spread you see in the tables below reflects that real-world variability rather than any tool overhead.

The benchmark files are public on GitHub. You can run them against whichever app you have installed and compare.

Recurring operations: exports and updates

#

These are the operations most stores run weekly or monthly, and they’re what your day-to-day experience with a bulk tool actually looks like.

Order exports

#

About 26 orders per second on the 1,000-record export. If you’re pulling orders for accounting close, period reporting, BI extracts, or a migration to another platform, this is what those jobs look like.

Product exports

#

The simplest version is just products and variants, no images or metafields. A 1,000-product catalog comes out in about a minute.

Most real exports include more than that. Pulling the full set of associated data (images, product metafields, and variant metafields) on the same 1,000-product catalog is a bigger job because each product carries more to read.

If you’re replicating this with Matrixify, make sure both product and variant metafields are selected in your export options, otherwise you’re timing a smaller job.

Bulk variant updates

#

This test takes a 1,000-product catalog and updates Variant Price and Variant Compare At Price across all 2,111 variants, keyed by Variant SKU.

About 11 variants per second. The same throughput pattern applies to other variant-level update operations: cost, SKU, barcode, weight, inventory.

This is the test most worth focusing on if you’re trying to picture what bulk work feels like in practice. Updates dominate real-world bulk activity (price changes, tag adjustments, inventory updates, status flips), and they hit different Shopify APIs than creates.

Migration-style imports

#

These are the bigger jobs, typically run once during a platform migration, an ERP reset, or a bulk historical-data load.

Order imports

#

Test orders use realistic line items, shipping and billing addresses, tax lines, and discount codes. Email addresses are on example.com, a domain reserved by RFC for testing.

Product imports (creates)

#

These tests create products in draft status with realistic field coverage: title, vendor, type, tags, body HTML, three variants per product, three images, and five metafields covering common types.

A note on the 1,000-product number. Roughly a quarter of that time is Shopify processing images on their side, not Altera doing the upload work. Re-running the same import with images stripped out (everything else identical) completed in 20m 30s. Image processing happens on Shopify’s infrastructure regardless of which import tool you use, and it scales with how many images each product carries.

Can Altera handle larger volumes? Yes. The biggest job in this published set is 1,000 records, but Altera has no per-job size limit on any plan, and we regularly see jobs in the hundreds of thousands. We’re publishing larger-volume benchmarks separately. If you have a specific scale you’d like numbers on before then, get in touch.

What shapes the numbers on your store

#

Bulk timing depends on more than the app pushing data. A few things shape what you’ll see on your own store.

Your Shopify plan. Non-Plus stores have a daily variant creation limit and tighter API rate limits, which can become the bottleneck on large first-time creates. Updates and deletes are usually fine.

Shopify’s load. Shopify’s APIs are shared infrastructure, so performance varies by time of day, region, and whatever else is happening on the platform that day. The min/max spread in the tables above reflects this.

The shape of your data. A 1,000-product file with one variant per product is a very different job from 1,000 products with 100 variants and 20 images each. We picked a realistic mix for these tests: most products have a single variant, ~12% have a small 2-5 size run, ~3% have a 10-30 size/color matrix, and a couple of outliers sit at 100-500 variants. Roughly 60% of products carry one image and 40% carry two, plus five product metafields and one variant metafield. Catalogs with more images per product will run proportionally slower, and catalogs without images will run faster.

What you’re doing. Creates, updates, and deletes hit different Shopify APIs. Updates are typically a lot faster than first-time creates, as the bulk update numbers above show. Deletes generally fall in between.

One thing that doesn’t tier: speed

#

Altera uses the same processing on every plan. Whatever plan you’re on, you get the same threading, the same throughput, and no per-job size caps.

This is worth saying because some import/export tools tier their performance by price. Cheaper plans run single-threaded or with fewer parallel workers; faster ones cost more. Altera doesn’t work that way. The plan you pick determines features and seats, not how fast your jobs run.

How does this compare to Matrixify?

#

We want to publish a full Altera vs Matrixify comparison, but we want to do it properly. That means the same Shopify plan, matched plan tiers, files run multiple times on each side, and results we can defend. That work takes time and we’d rather hold the post until it’s right than ship something half-done.

That said, in our informal testing we’ve seen that Altera compares very favorably. But we’d rather hand you the test files than ask you to take our word for it.

We’ve put a set of sample files on GitHub. Each file fits within Matrixify’s Basic plan limits, and the smallest one fits their free Demo plan. Run any of them on whichever app you have installed, time the result, and see what you get.

A note on pricing if you’re planning to compare. Altera Pro is $15/mo and runs the same threading and throughput as every other Altera plan. The closest Matrixify equivalent in capability is the Big plan at $50/mo, where their parallel threading kicks in. Their cheaper plans (Demo and Basic at $20/mo) are single-threaded, so results from those won’t reflect Matrixify’s parallel performance. The comparison most stores will want to run is Altera Pro vs Matrixify Big.

If you do run the files we’d love to hear what you saw. Your Shopify plan, the Matrixify tier you tested, and the times you measured. We’re collecting community data for a follow-up post.

Bottom line

#

If you’re considering Altera and the question is “can it handle our store”, yes. That includes catalogs and order histories well into the hundreds of thousands.

For typical recurring bulk work (price updates, tag changes, inventory adjustments, order exports for accounting), jobs finish in minutes. The slowest scenario in this published set is a 1,000-product first-time create with full image processing, which finishes in under half an hour, and most of that time is Shopify processing images on their side rather than the import tool.

If you have a scenario you want to test before committing (a particular file shape, a specific Shopify plan, an unusual data type), get in touch. We’d rather show you it works on your data than have you guess.