Procurement Intake and Orchestration Software: Build vs Buy

What the off-the-shelf products actually do well

Buy. For most companies reading this it is not a close call, and the products in this category are good enough that talking yourself into a build is an expensive kind of ambition.

Zip largely defined this category and its configuration engine covers a wide share of ordinary routing: thresholds, conditional reviewers, parallel approvals, service levels, a status page the requester can read. Levelpath stands up quickly and its intake experience is the kindest to an engineer who has never bought anything before. ORO Labs suits a team that wants to own process design rather than file change requests with a vendor. If you run Coupa, SAP Ariba Guided Buying or Workday Strategic Sourcing, intake is bundled, and you should use it for a year before deciding it does not fit.

What every one of them gives you free is worth counting. Maintained connectors. Someone else keeping single sign-on working when your identity provider changes. Someone else answering on a Monday when a reviewer cannot open a request.

If your approval policy fits on two pages, your thresholds are the same in every legal entity, and you run one Enterprise Resource Planning (ERP) system, buy one of these and spend the difference on category strategy. A custom build at that profile will lose on speed, on cost and on maintenance, and we say so on first calls with our own proposal sitting on the table.

The rest of this page is about the ceiling, and what it looks like from underneath.

Where they stop: the request that changes shape halfway through review

An engineering manager wants a monitoring tool. Forty two thousand dollars a year. She fills in a form she found on the intranet, which turns out to be last year's form. Procurement asks whether security has reviewed it. Security wants a completed vendor questionnaire and a SOC 2 Type II report. The vendor sends one that expired in March. Legal will not look at the order form until the data processing addendum is attached. Privacy wants to know whether personal data leaves the region, which nobody has asked. Tax raises a withholding question because the entity signing is not the entity paying.

Seven weeks later the purchase order issues, the renewal date the vendor quoted has moved, and she has quietly been running a smaller plan on her corporate card since week two.

Nobody in that story did anything wrong. Each function asked a reasonable question at the moment it saw the request. The failure is that the sequence and the criteria were never encoded anywhere, so the requester became the router, and the requester has no idea what the rules are.

Configuration engines handle a good share of that with conditional logic. Where they break is dependency and re-evaluation. A request that starts as software becomes a professional services engagement halfway through review, which retroactively changes who must approve. A privacy answer changes after security signed off, invalidating an approval already given. Most packaged tools model approvals as a linear chain and cannot reopen one completed step without restarting everything.

That single behaviour decides whether people use the tool or route around it. The tell is easy to spot: you bought an intake product and your buyers still keep a shadow tracker, because the product cannot express the exceptions your policy actually contains.

The second gap is memory. Three teams buy three overlapping observability tools in one year. A contract auto-renews because the notice window passed while the request sat in review. A supplier gets onboarded twice under two spellings. All three are the same failure: intake has no knowledge of the estate it is adding to.

The arithmetic: platform fees and requester counts versus a build

Intake and orchestration products price on some combination of annual spend under management, request volume tiers and integration count. Connectors into an ERP or a contract system are frequently separate line items, and the price of a second ERP is rarely proportionate to the first.

Here is the crossover as a number. Under roughly 1,500 intake requests a year, one ERP instance and one set of thresholds, buy. Between 1,500 and 6,000 requests with two entities, the packaged tool still wins, though expect real money on configuration services and a shadow tracker surviving somewhere. Above roughly 6,000 requests a year, or three or more ERP instances, or more than about 400 active requesters, a three year total for licence plus configuration plus the integration work you still fund yourself routinely crosses the full platform band below.

Then price the part no invoice covers. Pull your last 100 completed requests and split elapsed time two ways: waiting on the requester, and waiting on a reviewer. Most procurement operations leads guess the second is larger. In our experience the first usually is, because nobody told the requester what was needed until somebody asked for it.

Multiply the median cycle time by your annual request volume and by the hourly cost of the people waiting. That figure is what intake already costs you, without a supplier ever invoicing for it.

What a custom build actually costs

Bands, from Digital Heroes delivery experience across enterprise internal tooling. A focused first release, meaning one intake experience, the policy rules engine with recomputation, reviewer assignment with service levels, requester status visibility and ERP requisition writeback, runs $70,000 to $150,000 and ships in 10 to 16 weeks. A full platform adding contract system handoff, renewal intake generated from contract dates, supplier and duplicate detection, integrations into Jira and ServiceNow, document extraction, reviewer workload management and cycle time analytics runs $180,000 to $400,000 phased across 6 to 11 months.

Data migration adds 10 to 25 percent. In this category it is mostly supplier and contract records rather than transactions, and the expensive part is fuzzy matching supplier names so your duplicate detection has something true to check against.

Year two runs 15 to 20 percent of build cost annually, and unusually for internal software that spend is predictable, because it tracks policy change. Every threshold your general counsel revises is an edit.

What pushes the number up: the number of ERP instances, because SAP in one region and NetSuite in another means two requisition models and two chart of accounts mappings. Entitlement complexity behind single sign-on. The number of reviewer tools you must push work into. Multi entity tax and legal rules, the most detailed part of any policy and the part most likely to change mid build. And the policy itself, which frequently does not exist in writing.

What keeps it down: launch with software and services purchases only, where the pain concentrates, and leave direct materials and capital expenditure on the current path.

The four situations where building wins

• Regulatory fit. Multi entity groups carry obligations that packaged routing models badly. Withholding and permanent establishment questions turn on which entity signs versus which pays. Cross border personal data needs a data processing addendum and, out of the European Economic Area, standard contractual clauses with a transfer assessment. Sanctions and export control screening has to happen before an order is placed, not after. Each of those is a mechanical rule against request attributes, and each one is a place where a wrong default is a finding rather than a delay.
• Scale economics. Pricing that climbs with requester count at exactly the moment you want every employee able to raise a request, so procurement stops being a help desk and starts being a function.
• A workflow that is your competitive advantage. If your commercial edge is how quickly a business unit can safely buy something, the routing engine is the business rather than administration around it. That is also why the recomputation behaviour matters so much: a tool that restarts an approval chain when one answer changes trains your best buyers to work in email.
• Integration sprawl across three or more systems. The ERP, the contract lifecycle system, the security queue in Jira or ServiceNow, the vendor risk platform, the supplier master and the identity directory. Reviewers will not adopt a new inbox. Work has to be pushed into the tool each reviewer already opens every morning, with actions available in place, and status pulled back. That is real integration work, not notification emails with links.

Two of those true is a build. One of them is a better configuration of what you already own.

How to decide in a week

Take forty completed requests from the last quarter and read them properly. That is the whole test.

Monday and Tuesday: for each one, record the elapsed days, and split that time into waiting on the requester and waiting on a reviewer. Record how many distinct systems the request touched, and how many times somebody retyped the same data into a second one.

Wednesday: count the reopens. How many requests had a material attribute change after an approval had been given, and what happened. If the answer is that the chain restarted, or that somebody approved out of band and backfilled it later, write down which.

Thursday: ask each reviewer group one question. Where do you actually see this work, and would you open a procurement tool to do it. Write the answers down verbatim, including the impolite ones.

Friday: put three numbers on one page. Median cycle time, the share of it spent waiting on requesters, and the count of requests that were reopened or worked around. If cycle time is under two weeks and reopens are rare, configure what you own better and buy the connector you have been avoiding. If a third of your requests changed shape mid review, and two reviewer groups told you they will never log in, no amount of configuration reaches that.

What follows is a paid discovery phase rather than a proposal. Two to three weeks, fixed fee, producing a signed product requirements document covering your real thresholds by entity, the requirement recomputation model, the reviewer push integrations and acceptance criteria. You own that specification whoever builds it, and you can hand it to three firms and finally get comparable quotes.

Who we are wrong for: single entity companies on one ERP, anyone shopping purely on hourly rate, and anyone who wants software before legal, tax and security have written their thresholds down. Digital Heroes writes that requirements document before any code, with more than fifty specialists and India LLP, US LLC and UK LTD entities so intellectual property assigns under your own law. ShopScore, HeroCheckout and Section Vault are our own products, over 2,000 projects sit behind us, and you meet the named team before signing. We are listed on Clutch, Trustpilot, Fiverr Vetted Pro and D-U-N-S.

Book a 30-minute call with Digital Heroes and get a written plan and a fixed quote within 48 hours.