Sign up with Google
Sign in with Zoho
Sales wants a customer portal. Finance needs the invoicing tool replaced before the vendor ends support. Production has asked for machine data in the ERP, and the managing director would like a dashboard by the next board meeting. Your IT team is two people, one of whom is also the help desk. Every request is reasonable, none can be done at once, and whichever you choose first will disappoint someone.
Prioritisation in this situation is not a scoring exercise, although scoring helps. It is a way of making trade-offs visible and defensible so that decisions are made on purpose rather than by whoever argues most persistently. This article sets out a practical method for doing that when capacity is genuinely scarce.
Start by Knowing What Capacity You Actually Have
Most overloaded IT teams overestimate how much time is available for projects. Before ranking anything, work out how many hours per month are truly free after support requests, maintenance, security updates, user administration and the unplanned problems that arise in every business.
A rough approach is to look at the last three months. How much time went into keeping things running, and how much into change? If the answer is that nearly everything went into running, then the honest conclusion is that the project budget is small, and the queue of requests needs to be cut accordingly. Overcommitting leads to projects that drag on for years, which is worse than saying no early.
It also helps to count the capacity outside IT. Business teams must contribute time to requirements, testing and training. A project that needs ten hours a week from the sales department cannot start if sales is in its busiest period.
Collect Everything in One Place
Requests arrive by email, in corridor conversations and in meetings. Put all of them into a single list, however small. For each, record who asked, what problem it addresses, which process or customers are affected, and any fixed deadline, such as a licence expiry or a legal requirement.
Describing the problem instead of the requested solution is more useful than it sounds. "We need a new CRM" tells you little. "Sales staff spend about two hours a week reconciling customer data between two systems" tells you what is at stake and leaves room for a smaller fix.
This list also exposes duplicates. Several departments asking for different tools often have the same underlying need, and a single project may serve them all.
Judge Each Request on a Few Honest Criteria
Elaborate scoring models with a dozen weighted factors give a false sense of precision. A handful of questions is enough, provided they are answered candidly and, where possible, with figures from your own business.
Business value
What changes if this is done? Value can be saved time, fewer errors, faster response to customers, new revenue or reduced risk. Try to estimate the effect in terms people recognise, such as hours per week, number of customer complaints or days to produce a quotation. Be cautious about vendor claims and treat estimates as ranges, not exact numbers.
Risk and obligation
Some projects are not optional. A system that will lose security updates, a process affected by a change in tax or reporting requirements, or a data protection gap all carry consequences that grow over time. Your tax adviser or legal counsel can confirm which obligations apply to your business. Such items belong near the top of the list regardless of how exciting other requests seem.
Effort and complexity
How much work is it, including integration with existing systems and data clean-up? A rough size, such as days, weeks or months, is enough. Include the effort required from business staff and the cost of external help if needed. Beware of projects that look small but depend on poor-quality data or an old system with no documentation, since these tend to grow.
Dependencies
Some projects unlock others. Cleaning up customer data makes a later CRM migration or portal much easier. Replacing an unsupported database may be a precondition for anything else that touches that system. These sequencing effects often matter more than the standalone value of each item.
Strategic fit
Does the project support where the business is heading, or does it patch something you plan to retire? Investing heavily in a system that will be replaced within a year rarely makes sense, unless it is a necessary stopgap.
Turn Judgements Into a Visible Ranking
A simple two-by-two grid, with value on one axis and effort on the other, is often sufficient. High value and low effort come first. High value and high effort need planning and possibly splitting into stages. Low value and low effort can be done opportunistically when time allows. Low value and high effort are the candidates to decline.
After placing items, adjust for risk and dependencies, since those override the grid. Then discuss the result with the people who own the budget and the business areas. The aim is agreement on the order and on what will not be done this year. A clear "not now" with a reason is far more useful to a department than an indefinite "we'll see".
Write the outcome down and share it. A visible list reduces the number of repeated requests and makes it easier to defend decisions when someone senior asks why their project is not first.
Break Large Projects Into Stages
When capacity is limited, a large project is a liability, because it blocks the team for months and delivers nothing until the end. Look for ways to split it so that each stage delivers something useful. Instead of a full customer portal, begin with order status and invoice access for the ten largest customers. Instead of a company-wide ERP replacement, start with one module, or build interfaces first so that data flows correctly before anything else changes.
Staging also reduces the cost of mistakes. If the first stage shows that users want something different, you have lost weeks, not a year.
Decide What to Build, Buy or Delegate
Capacity limits affect not only the order of projects but also how they are carried out. Where standard software covers most of the need, buying and configuring is usually the better use of a small team than building from scratch. Where processes are unusual, tailored development may be justified, but it requires ongoing maintenance, which a small team must be able to sustain.
External help can extend capacity if it is directed at the right tasks. Businesses in northern Germany sometimes engage providers offering Softwareentwicklung in Hamburg for projects that need close coordination with operational teams on site. For defined pieces of work such as building interfaces, migrating data or developing a specific module, some companies add temporary capacity through Softwareentwicklung Outsourcing while their own staff keep ownership of requirements and acceptance testing.
Whichever route you choose, protect the knowledge inside the company. Insist on documentation, access to source code where applicable and a proper handover, so the business is not left depending on a supplier for every change.
Review Regularly and Protect Time for Maintenance
Priorities shift. A new customer requirement, a security problem or a change in regulation can reorder the list. Review it every quarter, and keep a small reserve of capacity for the unexpected, so that every surprise does not derail planned work.
Reserve time for housekeeping too: updates, documentation, retiring unused tools and reducing technical debt. These tasks rarely win in a value comparison because their benefit is the absence of future trouble, but neglecting them eventually consumes more capacity than they would have required.
For companies that need a bespoke application alongside their routine IT work, teams offering Softwareentwicklung in Berlin are another option to consider. As with any supplier, ask for comparable project examples, clear scope and defined support after delivery.
Conclusion
Good prioritisation with limited IT resources comes down to three habits: knowing your real capacity, judging requests on a few honest criteria, and saying no clearly to what will not fit. A ranking that everyone can see, with reasons attached, is worth more than a sophisticated model that nobody trusts.
A useful first step is to gather every open request into one list this week and add two columns: the problem it solves in plain terms, and a rough estimate of effort. Then sit down with the department heads for an hour and agree the top three for the next quarter, along with what is explicitly left out. That conversation alone usually resolves more conflict than any scoring method.
