Replatform or Optimize? A Decision Framework for Ageing Ecommerce Stores (2026)

Two merchants, same night.

The first is staring at a slow admin panel at 11pm. Order count for the day is 340. The page has taken 22 seconds to load. Checkout conversion dropped three percentage points last quarter and she cannot tell if it is the app sprawl, the payment gateway, or the platform version her vendor stopped supporting. She has an agency proposal on the desk. £180K to replatform in six months. She does not know if she is buying a solution or an escape.

The second merchant signed that same proposal last year. The new store launched on time and on budget. Six months in, conversion is up four points and the admin is fast. But churn is worse. Subscription retention has not moved. The checkout still drops, and the reasons look strangely familiar. He spent £180K to fix problems that turned out to be workflow problems, not platform problems.

Somewhere between those two nights is the right answer. This post is the framework for finding it.

Why This Decision Is So Often Wrong

We have watched the replatform-or-optimize decision go wrong in three predictable ways.

Pattern 1. Replatforming to fix what a workflow could fix. A mid-market fashion brand comes to us asking for a migration off Shopify Standard because "the checkout is broken." Two hours into the audit, the checkout is fine. Customer service is reversing a large share of orders because a Klaviyo automation is misconfigured. A new platform inherits the automation and the reversal rate. Same problem, bigger invoice.

Pattern 2. Optimizing when the platform genuinely cannot scale. The mirror-image mistake. A brand sinks a year of custom patches into a platform version that already hit end of life, buying six months of relief with each patch. Optimizing a platform that has no supported upgrade path is not optimization. It is denial with a purchase order.

Pattern 3. Ignoring the middle option. The framing "replatform or optimize" is itself part of the problem. Most stores between £5M and £50M do not need a full replatform. They need aggressive optimization, or a targeted headless build, or a partial migration where only the checkout or only one market moves. We ship this middle path more than either of the extremes, and it is almost never the option agencies present first.

Agencies quietly love replatforming because it is a bigger contract. Optimization billings are roughly one third the value. This is why you rarely see a genuinely balanced "should you replatform" post from an agency blog. We publish this one because we build both, and we would rather do the right thing than the bigger one.

The Five-Variable Decision Framework

The framework is a scoring matrix. Each variable is scored 1 to 5, where 1 pulls the decision toward optimization and 5 pulls it toward a replatform. Add the five scores. The total tells you what to do.

Variable 1. Revenue trajectory

The variable most operators skip because they focus on where they are, not where they are going. A store growing 40% year on year has fundamentally different platform needs from a store that is flat.

Score 5. Growing 30% or more YoY with a demand signal that will keep compounding, new markets, new categories, retail rollout, wholesale acceleration. Your current platform will be the bottleneck within 18 months.

Score 3. Growing 5 to 30% YoY. The platform is not urgent yet but it will be inside 24 to 36 months.

Score 1. Flat or declining. A replatform will not fix a demand problem. Fix the demand problem first, then re-evaluate.

The trap. "We are about to grow" is not trajectory. Ambition is not data. If the growth rate is not in the last 90 days of your own numbers, do not score it as trajectory.

Variable 2. Technical debt score

Technical debt is not one number, it is ten. Score your store against this checklist. One point per statement that is true.

• Your platform version is out of vendor support within 18 months.
• You have custom modules or extensions that break on every core upgrade.
• You have more than 20 apps installed with overlapping functionality.
• You have checkout customizations that block or delay platform upgrades. (Shopify merchants: see our guide to the August 26 checkout deadline for why deprecated checkout surfaces are a specific debt symptom.)
• Your mobile PageSpeed Insights score is under 40 on your product page.
• Your payment gateway or ERP integration is hitting API limits or timing out at peak.
• You need headless or partial-headless capability and your platform does not support it cleanly.
• Your multi-region or multi-language setup is running through workarounds, not native features.
• Your admin panel takes five seconds or more per action for your operations team.
• Your feature-ship velocity, measured in customer-facing features per quarter, has halved year on year.

Count your score.

Score 5. Eight or more are true. The platform is genuinely constraining the business.

Score 3. Four to seven true. Debt is real but selective. Optimization can fix the worst of it.

Score 1. Three or fewer are true. Your platform is not the problem. Something else is.

Variable 3. Total cost of ownership vs rebuild

The variable most decisions get emotional about. Get the math on the table.

The table below shows typical ranges we see across Huptech Web engagements and public case study benchmarks from Shopify Plus, Adobe Commerce implementation partners, and commercetools. Use these as a starting range only. Run your own numbers against your traffic, AOV, and baseline conversion rate before you decide anything.

Break-even is the month at which the incremental revenue from the new platform, coming from higher conversion, faster site, headless features, or better internationalization, exceeds the cost delta between optimization and replatform.

Score 5. Break-even inside 18 months. The math already supports the rebuild.

Score 3. Break-even 18 to 30 months. Marginal. Depends heavily on your growth rate.

Score 1. Break-even over 30 months. Do not replatform on the math alone.

Any agency that quotes you a break-even without asking for your baseline conversion rate, average order value, and traffic is guessing. Push back until they show their working.

Variable 4. Team capacity for a 90 to 180 day project

Replatforms fail on the client side more often than the industry admits. Score honestly.

• Do you have a project lead who can dedicate at least 40% of their time to the migration for six months?
• Do you have a QA function, in-house or contracted, that can absorb a full-site regression test?
• Do you have executive sponsorship, CEO or CFO or equivalent, actively backing the migration and available for weekly decisions?

Score 5. All three yes. You can absorb the project.

Score 3. Two yes. Manageable but you will slip a milestone or two.

Score 1. One or zero yes. Do not start a replatform. The project will consume the team and you will resent the new platform for problems the project created.

The failure mode is starting a replatform with borrowed capacity. Marketing loans a manager for 20% of their time. Ops loses their strongest analyst to a full-time project role. Neither team signed up for that scope. Twelve weeks in, marketing is angry, ops is understaffed, and the migration is behind. The platform did not cause any of that. Capacity did.

Variable 5. Platform vs workflow diagnosis

The most-missed variable. Before scoring, list the five problems you are trying to solve by replatforming. For each, ask the question the framework hinges on. Would this problem survive the replatform?

Example. "Our checkout drops 20% on mobile." Would a new platform fix that? Sometimes yes, a modern platform's checkout out-converts an older custom checkout at scale. Sometimes no, if the drop is caused by a popup timing issue, an address-form bug, or a shipping-rate misconfiguration, a new platform inherits all three.

Walk through your five problems honestly.

Score 5. Four or five of your five top problems are genuinely platform-caused. A new platform will resolve them structurally. Score 3. Two or three are platform-caused. Partial migration or aggressive optimization is likely a better fit than a full replatform. Score 1. One or zero are platform-caused. Do not replatform. Your problem lives in your team, your process, your app config, or your product.

This variable alone has saved several Huptech clients from significant replatform spend in the last 18 months.

Reading Your Score

Add the five scores. Total range is 5 to 25.

20 to 25. Replatform. The math and the technical debt both support it. Move.

15 to 19. Optimize deeply first. Fix the 30% of technical debt causing 70% of the pain, then re-score in six months. If the score jumps to 20+ after optimization, you have a genuine platform problem. If it stays put, optimization was the right call.

10 to 14. Optimize and monitor. You are not in the replatform zone yet. Ship the optimization playbook below, harden integrations, and put a 12-month review on the calendar.

Under 10. Do not replatform. Your problems are strategy, product, workflow, or demand. Replatforming a store scoring under 10 is expensive procrastination. Fix the actual problem.

Download the scoring spreadsheet at the end of this post to run the math on your own store.

What Optimization Actually Looks Like (When the Score Says Optimize)

Optimization is not "do nothing." It is a specific engagement type with defined outcomes. We ship six optimization workstreams in a typical 12-week engagement. Each has a measurable target. We describe the shape of each workstream below. Specific lift numbers vary by store baseline, so we describe outcomes qualitatively here and confirm ranges with clients against their own baseline before we start.

App audit and consolidation. Inventory every installed app. Score for overlap, performance impact, and revenue contribution. Consolidate or remove. The typical outcome is a faster admin panel, a leaner storefront, and material savings on redundant app subscriptions.

Theme performance rebuild. Rebuild the storefront theme against modern PageSpeed targets. Remove render-blocking scripts, defer non-critical assets, optimize image delivery. The typical outcome is a meaningful PageSpeed Insights lift on mobile, which typically improves mobile conversion. Checkout audit and CRO pass. Focus areas are shipping options, payment method ordering, express-checkout placement, and address-form logic. Ship one change at a time and measure against the baseline. The typical outcome is a checkout conversion lift without touching the platform.

Klaviyo (or Omnisend, Bloomreach) flow rebuild. Rebuild welcome, abandoned cart, browse abandonment, post-purchase, and win-back flows. Segment aggressively. The typical outcome is a lift in email revenue over a 90-day window.

Custom-app-to-Shopify-Function migration. Migrate custom logic living in third-party apps into native Shopify Functions where possible. The typical outcome is performance parity or better with a cleaner upgrade path and reduced app sprawl. See our Shopify Scripts to Functions migration guide for the specific playbook. This one is worth its own project on any Shopify Plus store carrying three or more custom apps.

ERP or OMS integration hardening. Rebuild fragile integrations against modern middleware patterns. The typical outcome is far fewer order-sync failures and much less manual reconciliation. Applies most to jewellery, fashion, and multi-warehouse operations.

What Replatforming Actually Looks Like (When the Score Says Replatform)

If your score is 20 or higher, here is the realistic timeline. Anyone who quotes you faster is either doing less work or hiding scope.

Phase 1. Discovery and architecture. Weeks 1 to 4. Data audit, catalogue mapping, integration mapping, custom logic documentation, SEO baseline (top 200 URLs by traffic and revenue), redirect matrix draft. This is where scope creep starts if the team is not disciplined. Under-invest here and the whole project drifts later.

Phase 2. Build. Weeks 5 to 14. Theme build (or headless front end), app installation and configuration, integration wiring, checkout customization via Shopify Functions or the equivalent. The staging environment stood up for stakeholder review by week 10.

Phase 3. Data migration and QA. Weeks 15 to 18. Product, customer, and order history migration. Full-site regression test. SEO migration, including URL mapping, redirect matrix implementation, structured data preservation, and internal link updates. Payment gateway and fulfilment integrations tested end to end with live test orders.

Phase 4. Launch and post-launch. Weeks 19 to 24. Cutover window, DNS switch, monitoring dashboards live, hotfix response plan active, first major peak event (Black Friday, Ramadan, category launch) on the new stack.

The honest bit. Add 30% to any migration timeline you are quoted. First-time replatformers usually miss the SEO migration scope and the ERP integration scope. Both are treated as a single line item in Phase 3. Both are actually multi-week workstreams.

The Three Migration Paths Nobody Talks About

Most posts frame replatforming as all-or-nothing. It is not. Three middle paths exist and fit the majority of stores between £5M and £50M.

Path 1. Checkout-only migration

Keep your current platform for catalogue and account management. Migrate only the checkout to a modern stack using a headless setup. Cost is roughly 40% of a full migration. Timeline is 8 to 12 weeks.

Fits stores where checkout conversion is the primary pain and the rest of the platform is functional.

Path 2. Catalogue-first migration

Rebuild the storefront on the new platform while keeping the legacy admin running in parallel for a defined period. Cost is roughly 60% of a full migration. Timeline is 12 to 16 weeks.

Fits stores where the customer experience needs a step change but back-office data (product, order, customer, ERP) is complex enough to warrant a slower migration. Common pattern for jewellery, luxury, and B2B-heavy stores.

Path 3. Multi-market split

Replatform your fastest-growing market first (typically UAE, US, or a specific EU country). Leave other markets on the legacy stack. Migrate the rest over 12 to 18 months as the ROI on the pilot market proves the case.

Fits multi-region brands with clear growth-market prioritization. This is the path we most commonly ship for UK brands entering the GCC and for US brands entering the UK and EU.

Which path fits which score profile.

Platform-Specific Sub-Decisions

For readers already at 20+ who are choosing where to replatform to. We ship on Shopify Plus every week, so we have a bias, and we will name it. We also recommend against Shopify Plus when the fit is wrong. Here is the practitioner view for the four platforms that come up most.

Shopify Plus if...

• You are under £200M GMV
• Your business is DTC-heavy, roughly 60% or more of revenue
• Speed to launch matters, you want to be live in five to six months, not twelve
• Your team wants managed infrastructure rather than owning uptime and hosting
• You want headless as an option but not a requirement

BigCommerce Enterprise if...

• Your B2B pricing complexity genuinely exceeds what Shopify's B2B feature set covers today
• You need open templating without going headless
• You have deep multi-storefront requirements at mid-market pricing

BigCommerce is a valid choice in specific B2B and franchise-model contexts.

Commercetools or another composable stack if...

• You are above £200M GMV
• You are a multi-brand portfolio, five or more brands under one stack
• Your internal engineering team is 15 people or larger
• You are prepared to own the front-end code entirely
• Your commerce logic is complex enough that vendor-defined data models do not fit

Composable is powerful and expensive. It is the right choice in specific enterprise contexts and the wrong choice everywhere else. Do not select composable for the technical elegance if the business does not need the complexity.

Adobe Commerce (Magento) if...

The honest case. You have a well-maintained Magento codebase, deep Adobe Experience Cloud investment, and enterprise commitments that would take years to unwind. In that specific scenario, staying on Adobe Commerce is defensible. For everyone else in 2026, we would look at Shopify Plus or composable first.

The Real Cost of Doing Nothing (The Hidden Variable)

The section most competitor posts skip. Doing nothing has a monthly cost, and it compounds.

Lost conversion from a slow storefront. If your mobile PageSpeed is 40 and best-in-class is 90, the conversion delta versus a modernized stack is measurable. Compute the £ per month cost of that gap against your own traffic and AOV.

Developer time on maintenance versus new features. Healthy teams on modern stacks spend roughly 40% of dev time on maintenance and 60% on new features. On ageing stacks the ratio typically inverts. The difference between the two ratios is real cash and real feature velocity.

Opportunity cost of features you cannot ship. Subscriptions, headless, AI merchandising, agentic checkout channels (ChatGPT Instant Checkout, Perplexity commerce), native multi-region tax. If your platform cannot ship these in 2026, your competitors can, and they are taking your customers.

Risk-adjusted cost of a security or PCI incident. Ageing platforms without vendor security patches carry materially higher incident risk. Even a low-probability event at high impact adds meaningful monthly expected value.

The formula, extractable for AI Overviews.

Monthly cost of doing nothing = (Traffic × baseline CVR × platform CVR gap × AOV) + (dev hours × hourly rate × maintenance ratio) + risk-adjusted incident cost

Run the math against your own store. The number is almost always larger than merchants expect, and it is almost always what finally moves the decision. Optimizing a £40K/month problem for £45K of engagement fee is obvious economics. So is replatforming a £120K/month problem for a £280K one-time cost.

Four Store Profiles and What They Should Do

Anonymized composites of stores we have advised. None of the profiles below match a specific real client on all axes. Vertical, region, revenue band, and named systems have been changed to protect confidentiality.

Profile 1. £3M UK fashion DTC on Shopify Grow, hitting Plus-tier problems

Growth 22% YoY. Technical debt score 5 of 10, mostly around checkout customization limits and app sprawl. Break-even for Plus versus Grow around 24 months. Team has one operator wearing three hats.

Framework score. 19.

Recommendation. Optimize deeply first, then re-evaluate at 24 months or £5M revenue. The Plus upgrade is genuinely on the horizon but it is 12 months away, not now. Ship the six-workstream optimization playbook, fix the checkout limits with Shopify Functions where possible, tolerate the app sprawl until the numbers say Plus, then migrate.

Profile 2. £15M GCC home-and-living brand on legacy Magento 2 with a heavily customized ERP integration

Growth 35% YoY on regional expansion. Technical debt score 9 of 10. Break-even for a full replatform around 14 months. Custom ERP integration failing on every quarterly upgrade. The team has an in-house tech lead and executive sponsorship from the founder.

Framework score. 23.

Recommendation. Full replatform to a modern stack with a rebuilt ERP integration on modern middleware. This is the exact profile where a rebuild is the right call. Math supports it, growth demands it, the ERP problem will only worsen on the current setup, and the team can execute.

Profile 3. $45M US supplement brand on Shopify Plus, hitting subscription-app limits

Growth 18% YoY. Technical debt score 4 of 10, concentrated in one custom subscription app that has become brittle. Break-even for a composable rebuild around 34 months. The team has strong dev capacity but is stretched.

Framework score. 12.

Recommendation. Do not replatform. Migrate the custom subscription logic to Shopify Functions and to a native Recharge or Skio configuration. Save the six-figure rebuild. The store is not the problem, the subscription app is. Fix the subscription app.

Profile 4. £200M multi-brand luxury group on an ageing composable stack

Growth 12% YoY across the portfolio, variable by brand. Technical debt score 7 of 10. Break-even for a per-brand Shopify Plus rebuild varies. Portfolio of six brands, two on healthy trajectories.

Framework score. 21.

Recommendation. Rebuild the fastest-growing brand as a Shopify Plus pilot in the next 12 weeks. Prove the case on one brand, gather real data, then decide on the remaining five over the following 18 months. Do not attempt a portfolio-wide replatform in one project. That is a common failure pattern at this scale. Multi-market split (Path 3 above) applied to a multi-brand portfolio is the right sequence.

Six Questions to Ask Any Agency Pitching You a Replatform

Arms you against sales-driven replatform pitches.

1. What score does our store get on your framework? Show us the matrix. If they do not have a framework, if they cannot articulate variables, or if the answer is "yeah you probably need to replatform," walk. Every reputable agency in 2026 runs a decision framework before quoting a rebuild.

2. What is the total 24-month cost of ownership, not just the build cost? Build cost is typically 40 to 60% of the two-year total. Ongoing platform fees, app costs, ongoing engineering, and post-launch optimization make up the rest. Any quote that only covers the build is understating the real spend.

3. What percentage of our current problems will survive the migration? A good agency will tell you honestly. Some problems are workflow, some are product, some are team. A replatform fixes none of those. If the agency cannot separate platform problems from non-platform problems, they will not fix either.

4. Have you shipped a store of our size on this platform in the last 12 months? Show us. Named client, named platform, named numbers. Case studies from 2019 do not count in 2026. Migration playbooks evolve fast.

5. Who owns the SEO migration scope, and what is your redirect matrix approach? The single most under-scoped area of every replatform. If the answer is vague, expect an organic-traffic drop of 20 to 60% at launch. The right answer includes a URL-mapping methodology, a redirect matrix template, a Google Search Console re-verification plan, and a structured-data preservation approach.

6. What is your rollback plan if the launch goes wrong in the first week? Every experienced migration team has one. Every rushed team does not. If the answer is "we do not need one, our launches do not fail," find a different agency.

If the agency cannot answer all six with real specifics, they are selling you a migration, not solving your problem.

When Optimization Fails and You Should Have Replatformed

The other honest section. Warning signs that optimization is stalling tactic, not strategy.

• Platform version is deprecated within 12 months and there is no supported upgrade path
• Cumulative optimization spend over the last 18 months is approaching what a full replatform would have cost
• New customer-facing feature launches are being blocked by platform constraints, not team capacity
• Third consecutive optimization project has failed to move the target KPI

If two of these are true, stop optimizing and replatform. Optimization was the right answer at some earlier point. It is not the right answer any more.

How Huptech Web Approaches This Decision

We ship both sides of the framework.

Aggressive optimization for stores where the score says optimize. Twelve-week engagements built around the six workstreams above. See our full performance optimization scope.

Full replatforms for stores where the score says replatform. Shopify Plus migrations shipped across USA, UK, and UAE markets. Details on our platform migration services.

Partial migrations (checkout-only, catalogue-first, multi-market split) for the middle-band stores that most agencies overlook.

Every engagement starts with the scoring matrix. We and the client are aligned on whether we are solving the right problem before we quote the wrong solution. If your score says optimize and we quote a rebuild, that is on us.

Two ways to move forward.

Download the scoring spreadsheet. Auto-calculates your score from the five variables and generates the recommended action, plus indicative cost bands.

Book a 30-minute decision consultation at /consult/platform-migration-services. Bring your current numbers (revenue, platform, growth rate, top five problems) and we will run the score with you.

Frequently Asked Questions

Q. Should I replatform my Shopify store?

Run the five-variable score. Replatform if the total is 20 or higher. Optimize deeply if 15 to 19. Optimize and monitor if 10 to 14. Do not replatform if under 10. The question is not "when will I outgrow the platform," it is "what is my score today across the five variables above."

Q. How long does an ecommerce replatform take?

A realistic Shopify Plus replatform from Magento or a custom platform takes 20 to 32 weeks end to end. Discovery is 4 weeks. Build is 10 weeks. Data migration and QA is 4 weeks. Launch and hyper-care is 4 to 6 weeks. Add 30% to any timeline you are quoted. First-time replatformers consistently under-scope SEO migration and ERP integration.

Q. How much does replatforming to Shopify Plus cost?

Typical ranges by GMV. £500K to £2M store, £45K to £110K one-time build. £2M to £10M, £110K to £280K. £10M to £50M, £280K to £800K. £50M+, £800K to £2.5M and above. Ongoing Shopify Plus platform fees start at $2,300 per month plus revenue share above certain thresholds. Add 12 months of post-launch optimization to compute total cost of ownership.

Q. What are the signs I need to replatform?

Six clear signals. Platform version deprecated within 18 months. Mobile PageSpeed under 40 with no fix path. Eight or more of the 10 technical debt statements above are true. Growth rate 30% or more YoY. Payment or ERP integrations hitting API limits. Feature ship velocity halved YoY. Two or more together indicate a genuine platform ceiling.

Q. Is replatforming worth it?

Worth it if the framework score is 20 or higher. Break-even is typically 12 to 24 months depending on store size. Not worth it if the score is under 15, break-even is beyond 30 months, or if the problems you are solving are workflow or product problems that a new platform will inherit. Around one in three "we should replatform" pitches we audit are actually optimization problems.

Q. What is the difference between replatforming and redesigning?

Replatforming changes the underlying commerce platform, for example Magento to Shopify Plus or custom to composable. Redesigning changes the front-end experience while keeping the same platform. Redesigns cost 15 to 40% of what replatforms cost and take 6 to 12 weeks versus 20 to 32 weeks. If your platform is fine but your storefront looks dated, redesign, do not replatform.

Q. Can I replatform without losing SEO?

Yes, if SEO migration is scoped from the start. That means URL mapping, a redirect matrix covering the top 500 pages by traffic and revenue, structured data preservation, sitemap regeneration, Google Search Console re-verification, and a 12-week post-launch monitoring plan. Done well, organic traffic recovers within 60 to 90 days. Done poorly, expect a 20 to 60% traffic drop that can take a year to recover.

Q. Should I move from Magento to Shopify Plus?

Sometimes yes, sometimes no. It depends on your Magento version support timeline, your B2B needs, your team capacity, and your total cost of ownership over 24 months. We have a dedicated deep-dive on this exact question with a Magento-specific decision framework at Should you move from Magento to Shopify?.

Q. What is the risk of replatforming?

Three main risks. SEO traffic loss at launch if migration is under-scoped. Team burnout if capacity was not honestly assessed. Post-launch discovery that the problems you were solving were workflow problems, not platform problems, meaning you spent six figures without moving the KPI. All three are avoidable with the framework and honest scoping.

Q. How do I know if optimisation is enough?

Optimisation is enough if your framework score is 14 or lower, your platform version has more than 24 months of vendor support, and no more than three of your top five problems are platform-caused. If those three conditions hold, ship the six-workstream playbook and re-score in 12 months. If the score has not moved by then, revisit the replatform question.

Q. What is a checkout-only migration?

A partial migration where you keep your current commerce platform for catalogue, admin, and back office, but move only the checkout experience onto a modern stack (typically headless on Shopify Plus). Cost is around 40% of a full replatform and timeline is 8 to 12 weeks. Fits stores where checkout is the dominant pain point and the rest of the platform is functional.

Q. When should a Shopify Plus store consider a headless rebuild instead of replatforming?

When you are already on Shopify Plus and the constraint is front-end performance, custom content architecture, or brand-experience differentiation, headless via Hydrogen or a custom Next.js front end is a better fit than moving to another platform. Cost is 30 to 60% of a full platform replatform. Timeline is 12 to 20 weeks. Keeps your commerce data, checkout, and admin intact.

Closing

Replatform when the math and the technical debt both say so, and only then. Optimize aggressively when either variable says wait, and use the six-workstream playbook to make optimization a real project, not a delay tactic. Consider the middle-path migrations (checkout-only, catalogue-first, multi-market split) before you accept the binary. And do not confuse "we need a bigger contract" (agency incentive) with "you need a bigger project" (your business need). If in doubt, run the score.

Book a decision consultation → /consult/platform-migration-services