Book a scoping call
Pricing, search and recommendations

Ecommerce software development: the engines that get slow first

The parts of a commerce stack that become engineering problems: pricing that recalculates in real time, search over a catalogue too big for naive queries, recommendations you want to own rather than rent, and the integration layer that quietly became the most fragile thing you run.

Book a scoping callRead the Recostream case study

Ecommerce systems we built, measured in production

Acquired
Recostream, the recommendation engine we built, bought by GetResponseDecember 2022
20 to 30 ms
to return a new recommendation under high event volumesRecostream
Intraday
repricing on every channel from live market feedsMennica Skarbowa omnichannel platform
Goals

What commerce teams want from their platform

  • Protect margin as prices move

    Prices that follow inputs through the day without holding up the page.

  • Help customers find the product

    Search that understands the words customers use.

  • Own your personalization

    Recommendations built from your customer record, on your terms.

  • One truth for stock and price

    Storefront, marketplace and shop floor agreeing on what is available and what it costs.

Challenges

Where commerce becomes an engineering problem

  • A platform that stops at your rules

    Pricing, bundling or merchandising logic the storefront platform cannot express.

  • Integrations that became fragile

    ERP, warehouse, marketplace and storefront connected point to point, each change a risk.

  • Search that misses

    Relevance and vocabulary problems that a faster engine will not fix.

  • Repricing that lags the market

    Price updates that arrive after the market has moved.

What we build

Ecommerce software we deliver

  • Real-time pricing engines

    Price and margin control that recalculates as inputs change, without holding up the page.

    Pricing engine case study

  • Recommendation engines

    Personalization from a customer record rather than the last item viewed. Recostream, the engine we built in-house, was acquired by GetResponse in 2022.

    AI Integration for Java Systems

  • Custom site search

    Search that understands your catalogue, including the synonyms your customers use and your product data does not.

  • Payment integration

    Gateways and processors, including the failure and reconciliation paths.

  • Systems integration

    Making ERP, warehouse, marketplace and storefront agree on stock and price, and connecting production systems such as label printing so orders flow without manual handling.

    Integration case study

  • Omnichannel platforms

    One set of rules serving storefront, marketplace and in-store rather than three that drift apart.

Case studies

Commerce systems we built

All case studies

  • The bespoke metal exchange platform works great, it is easily accessible and richly functional. Stratoflow managed deadlines capably, meticulously documented their progress, and delivered a complex project at an affordable cost.
    Bartłomiej KnichnickiVice Chairman, Supervisory Board, Mennica Skarbowa
How we help

Services for commerce platforms

When to use an agency

When a platform agency is the better choice

If you are running a standard catalogue on a mainstream platform, an agency that specialises in that platform will be cheaper and faster, and we will tell you so.

Questions we get asked

Ecommerce software questions

Do you build stores on Shopify, Magento or similar?

Not as our main work. We build the systems around and underneath them: pricing, search, recommendations, integrations and data. If your project is mostly storefront build and theming, a platform specialist is the better and cheaper choice.

Can you improve search on our existing catalogue?

Usually yes, and it tends to be the highest-return change available. Most search complaints are about relevance and vocabulary rather than speed, which means the fix is in how the catalogue is indexed and enriched rather than in buying a faster engine.

Should we build or buy a recommendation engine?

Buy first. A hosted engine will beat a hand-built one until your catalogue or your merchandising rules are unusual enough that the hosted model cannot express them. We built Recostream because that point does arrive, but it arrives later than most teams assume.

How do you handle stock and price consistency across channels?

By deciding which system is authoritative for each fact and how stale the others may be, then designing to that rather than trying to make everything instantly consistent. Most channel drift comes from having no explicit answer to that question.

You are a Java company. Do you know ecommerce?

We built Recostream, a recommendation engine for online stores, and sold it to GetResponse; we built the real-time pricing engine for Mennica Skarbowa, an omnichannel precious-metals business; and we have integrated order flows through to production systems such as label printing.

What technology do you use for commerce engines?

Java services for pricing and integration, Elasticsearch for catalogue search, the Java and machine-learning stack we built Recostream on for recommendations, and MuleSoft or Spring Integration where order flows cross systems.

Go deeper

Ecommerce engineering notes

Written by , Co-founder and CTO, Stratoflow. Published . Updated .

Next step

Book a scoping call

Thirty minutes with an architect, not a salesperson. You leave with a written view of scope, price model and whether we are the right team for it.

Book a scoping callSee the proof first