The 7 questions you must answer before investing a single dollar in technology (AI, CRM, automation)
Business · · by Gabriel Bertagnolli
Series: Decisions that move the needle: Article 2 of 10
Every week a salesperson appears with the tool that is "going to transform your company." The demo is flawless, the price has a discount if you sign this month, and the underlying argument is always the same: "if you don't adopt it now, you'll be left out." Under that pressure, thousands of SMBs buy software with the same logic as buying an online course: out of fear of missing out, not because of a business decision.
The result is documented and uncomfortable. Large-system implementations systematically fail to achieve their objectives: Gartner estimates that more than 70% of recently implemented ERP initiatives will not fully achieve their original business use case goals (Gartner, ERP Insights). McKinsey has pointed out for years that around 70% of large-scale transformations fail (McKinsey, 2022). The classic CHAOS report from The Standish Group shows that only 31% of software projects end in complete success, while 50% are "challenged" (with scope or budget cuts) and 19% fail outright (Standish Group, CHAOS Report).
And it is not just large software. In digital marketing, technology stack (martech) utilization dropped to 33% in Gartner's 2023 measurement, down from 58% in 2020 and 42% in 2022 (chiefmartec on Gartner, 2023); the 2025 measurement barely recovered to 49% (Gartner, 2025). In other words: out of every dollar spent on licenses, more than half funds tools that no one fully uses. Capterra, for its part, found that 60% of software buyers regret at least one purchase made in the past 12 to 18 months (Capterra, 2024).
The takeaway is not "do not invest in technology." It is that investment does not begin with the salesperson's demo: it begins with seven questions you need to answer before moving a single dollar. If one of them remains without a concrete answer, you are not yet facing a technology decision; you are facing an impulse.
The 7 Questions
1. What concrete business problem does it solve, and how would you measure it?
"Modernizing," "digitizing the business," or "adopting AI" are not problems: they are categories. A concrete problem has an owner, has a process that currently fails, and has a metric that is currently broken. The implementation question is: what number will I look at three months after rollout to know whether it worked? If you cannot name that number before buying, the software will not create it.
2. Is the problem about technology, process, or people?
This question saves more money than any price negotiation. If the sales process is undefined, a CRM will only log the mess with higher precision. If no one follows up on leads today in a spreadsheet, they will not do it tomorrow in a more sophisticated system. Evidence on implementation failures consistently points to adoption and change management factors, not software defects. Rule of thumb: if the problem persists when you simulate it manually in a spreadsheet, buying software only accelerates it; it does not solve it.
3. What is the total cost of ownership over 12 months, not the license price?
The license is the tip of the iceberg. The real cost includes data migration, training, internal setup hours, integrations with what you already have, and the cost of changing your mind if the tool does not work. An honest comparison is between the 12-month total cost of the tool versus the 12-month total cost of the current state (including the hours the problem costs you today). That delta is the investment to justify, not the list price.
4. What happens to our operations if the project is delayed or abandoned halfway through?
With 70% of ERP projects failing to meet their original goals and 50% of software projects "challenged" according to CHAOS, planning for partial failure is not pessimism: it is diligence. Ask yourself what keeps running if the project is frozen in month three: do you revert to the previous state unharmed, or are you worse off than before? Projects that degrade operations if abandoned halfway through demand more controls than those that allow a clean rollback.
5. Who within the company is the adoption owner?
Not the project owner: the adoption owner. They are different things. The project owner configures and trains; the adoption owner is someone with authority over the changing process, who is accountable for the metric from point 1, and who has dedicated time to sustain the change over the first 8 to 12 weeks. If the answer is "no one in particular, the vendor will handle it," put away the credit card: utilization will end up similar to the industry average (33% in 2023, 49% in 2025 in martech, according to the Gartner measurements cited above).
6. Can it be tested on a small scale before committing the entire company?
A well-designed pilot answers questions 1, 2, and 5 with limited spend: one team, one process, 30 to 60 days, and success criteria defined before starting. The uncomfortable question for the salesperson is what happens if the pilot fails: how data is recovered, what is lost, what contractual restrictions exist. A mature vendor has an answer for that. If the only exit is "the pilot converts into an annual contract," the pilot is not a pilot.
7. What is the cost of doing nothing: for a year?
Every salesperson assumes that the current state is free and that the tool is the only cost. Look at the opposite mirror: how much does one more year with the current process cost you? Hours of rework, errors from re-entering data, lost opportunities due to slowness. If that cost is real and growing, the investment is justified faster. If you cannot quantify or approximate it, the pressure to buy likely comes from a trend, not a problem. (A contextual data point: marketing budgets have already stalled at 7.7% of company revenue and 59% of CMOs report not having enough budget to execute their strategy according to Gartner (Gartner, 2025): money for "another tool" does not appear on its own.)
Decision Checklist (to print and fill out)
| # | Question | Concrete answer (one line) | Ready? |
|---|---|---|---|
| 1 | Problem + success metric | E.g.: "reduces billing cycle by 3 days; I track average days" | ☐ |
| 2 | Tech/process/people diagnosis | E.g.: "undefined process → first we define workflow" | ☐ |
| 3 | 12-month total cost vs. current state | E.g.: "US$9,600 vs. US$18,000 in hours today" | ☐ |
| 4 | Rollback plan if abandoned | E.g.: "weekly CSV export; cleanly reverts to spreadsheet" | ☐ |
| 5 | Adoption owner with name and time | E.g.: "commercial manager, 4 h/week for 12 weeks" | ☐ |
| 6 | Scoped pilot with upfront criteria | E.g.: "1 team, 45 days, goal: 80% weekly usage" | ☐ |
| 7 | Cost of doing nothing for 1 year | E.g.: "≈200 h of rework + 2 lost clients" | ☐ |
Cutoff rule: if 2 or more boxes are empty, the decision is not mature. It is not a definitive no; it is a "not yet." Return to the empty box before signing.
Three Red Flags in the Conversation with the Salesperson
- Artificial urgency. "The discount expires Friday" is a tactic, not a reason. An investment decided in 48 hours without a pilot or an adoption owner is exactly the purchase profile that feeds the 60% regret rate measured by Capterra.
- Demo metrics instead of your metrics. If the salesperson only shows product data and cannot help you define your success metric (question 1), they are selling you software, not an outcome.
- "All your competitors are already using it." Sometimes it is true. It still answers none of the 7 questions. FOMO is not an investment framework: it is the absence of one.
The Correct Order of the Decision
Investing in technology is one of the highest-return decisions available to an SMB, when done in the correct order: first the problem and its metric, then the process, then people and their adoption, and only at the very end the tool. The statistics cited above show what happens when that order is reversed: powerful tools, minimal utilization, abandoned projects, and budgets that fall short. The 7 questions do not hold back innovation; they make it fundable.
Sources consulted
- Gartner. Latest Enterprise Resource Planning (ERP) Insights. Accessed October 14, 2026: https://www.gartner.com/en/information-technology/topics/enterprise-resource-planning
- McKinsey & Company (2022). Common pitfalls in transformations: A conversation with Jon Garcia. Accessed October 14, 2026: https://www.mckinsey.com/capabilities/transformation/our-insights/common-pitfalls-in-transformations-a-conversation-with-jon-garcia
- The Standish Group. CHAOS Report on IT Project Outcomes. Accessed October 14, 2026: https://opencommons.org/CHAOS_Report_on_IT_Project_Outcomes
- chiefmartec (Brinker, S.) (2023). Martech utilization problems: how to diagnose and remedy them (Gartner 2023 Martech Survey data). Accessed October 14, 2026: https://chiefmartec.com/2023/08/martech-utilization-problems-how-to-diagnose-and-remedy-them
- Gartner (2025). Marketing Technology (Martech). Accessed October 14, 2026: https://www.gartner.com/en/marketing/topics/marketing-technology
- Capterra (2024). Tech Trends Report: Software Purchase Regret. Accessed October 14, 2026: https://www.capterra.com/resources/tech-trends-software-purchase-regret
- Gartner (2025). 2025 CMO Spend Survey (press release, May 12, 2025). Accessed October 14, 2026: https://www.gartner.com/en/newsroom/press-releases/2025-05-12-gartner-2025-cmo-spend-survey-reveals-marketing-budgets-have-flatlined-at-seven-percent-of-overall-company-revenue
